PDA

View Full Version : PSU Assistance



CNCStudent
04-08-2015, 09:12 AM
Hello, I've got a 42V 720W switched mode power supply running 4x AM882H stepper drivers with 4 x 60BYGH301B (http://cnc4you.co.uk/resources/Stepper%20Motor%20Nema%2023%2060BYGH301B%203.1Nm.p df)stepper motors (3.1nm/4.2A/0.65Ohm/parallel wired), running a 4x4 plasma table (dual motors to one axis) on a HTD belt drive acting as a rack & pinion arrangement.

It's been running fine, but not happy about the switched mode power supply really, and it was only really a temporary fix, hence I was looking at using a torodial transformer, given these are AM88H drives, they accept accept AC or DC as they have their own rectifiers and capacitors internal.

question, what voltage and VA rating would you go with...

I was considering 2x50V / 1000VA, but the choices are really 2x45V, 2x50V, 2x55V, 1000VA, 1250VA 1500VA, 2000VA.

from somewhere like here: (http://www.airlinktransformers.com/chassis_mounting_toroidal_transformers/chassis_mounting_toroidal_transformers_standard_ra nge/?filter_input_voltage=230&filter_output_voltage=50%2B50&filter_va=1000&submit=Filter+Results) / maplin do up to 1000VA.

Yup, I've read all the gecko drive power supply selection info, and I know all about ac > dc x sqrt 2, etc etc.

I am hoping to use the same drive box in different applications in my garage as I'm not intending to build another, hence for a lathe and milling machine if I ever get around to converting any of those.

The BOB PSU is separate

thanks for your time and any guidance.

Clive S
04-08-2015, 10:13 AM
I think you could run that off 500Va and use 2x50V using 1000Va you will have problems with inrush etc

Robin Hewitt
04-08-2015, 10:36 AM
Amazing, I would go in exactly the opposite direction and replace toroidal with switch mode :beer:

Transformers make everything squiffy, heavy and complicated.

Clive S
04-08-2015, 10:57 AM
Amazing, I would go in exactly the opposite direction and replace toroidal with switch mode :beer:

Transformers make everything squiffy, heavy and complicated.

Well from the drive info:-

Regulated or Unregulated Power Supply

Both regulated and unregulated power supplies can be used to supply the drive. However,

unregulated power supplies are preferred due to their ability to withstand current surge. If regulated

power supplies (such as most switching supplies.) are indeed used, it is important to have large

current output rating to avoid problems like current clamp, for example using 4A supply for 3A

motor-drive operation. On the other hand, if unregulated supply is used, one may use a power supply

of lower current rating than that of motor (typically 50%
~70% of motor current). The reason is that

the drive draws current from the power supply capacitor of the unregulated supply only during the

ON duration of the PWM cycle, but not during the OFF duration. Therefore, the average current

withdrawn from power supply is considerably less than motor current. For example, two 3A motors

can be well supplied by one power supply of 4A rating.

So take your choice:apathy:

Robin Hewitt
04-08-2015, 12:09 PM
So take your choice


I always do :thumsup:

I went off transformers for CNC when I tried to calculate the voltage to buy and what value for the smoothing cap.

CNCStudent
04-08-2015, 12:22 PM
The am882h drives have built in bridge rectifiers and caps, but when I had the cover off (I was curious they were sold as refurbished) I should have written down the capacitance of the caps, hence the ripple can then be calculated.

My switch mode is 720w at 42v and it can make a bit of noise when all three axis are running at the same time. Even though in theory it's rated at 14amps.

Not so sure about 500va given its 250va per tapping (5amp) and the drives run at 4.2a hence if two axis are running off the same tapping at the same time they would draw worst case 8.2a, hence why I was edging toward 1000va (500 va per tapping or 10a rating at 50v)

Thanks for the comments (any comments)

Brad

Jonathan
04-08-2015, 01:53 PM
I'm currently a fan of transformers as they're easy to fish out of the university skip...but when the price is comparable a switch mode power supply is sensible. You can always add components if required to increase the peak rating.


Not so sure about 500va given its 250va per tapping (5amp) and the drives run at 4.2a hence if two axis are running off the same tapping at the same time they would draw worst case 8.2a, hence why I was edging toward 1000va (500 va per tapping or 10a rating at 50v)

Be careful not to mix up motor phase current with DC-bus current - they are rarely the same.

Clive S
04-08-2015, 06:21 PM
With the switch mode you generate RFI I run 4 of those drives with 3.1nM Motors from a 500VA train with no issues like a lot of people on hear do.

CNCStudent
04-08-2015, 09:38 PM
transformer rated at 50v ac, once rectified will provide ~70v DC (with no load), is this not a little high given the geckodrive guidance is 32 x SQRT (L), where L is in mH.... [32xSQRT(3.2) = 57v]

there is also another bit of guidance that says, no more than 20 x the stepper DC rated voltage (4.2A x 0.65 Ohms = 2.73v; 2.7v x 20 = ~54v), or no more than 25x the dc rated voltage or the steppers would overheat at standstill.

I am aware of ripple voltage will have onto the power supply when under current draw (hence interested to know what the caps are in the AM882H's)...

other than cost... what would be the problem with a 1000VA transformer, inrush on a 13A fused plugtop on a 32A ringmain should not be an issue, and if it was, I could always use a solid state relay which would then allow for switching only a the crossover voltage.

I got the 42V switch mode PSU, and although the nameplate says 720W (17A), it should really be able to deal with 4 stepper motors really, or at least just about (hence I was going to put one of the axis on another switchmode PSU), but it just seems to be like putting a plaster on the problem instead of getting the right power supply (a toroidal transformer) with a fair bit of design margin.

Lets face it, I have a welder that is 200A rated for 6mm mild steel, but only ever tend to use it at ~110A on 2 and 3mm plate, is it really that detrimental having a bit of margin in components? [ok, it is inverter driven, digital controlled... semi-synergic...]

Interesting in that I thought the guidance would come on voltage choice and not a VA rating discussion... as I could not make my mind up if 50V or 55V would be right for the steppers and drives or whether I was over rating it and should have been looking at 45v transformer output rating, given the no load DC voltage would be about 63V.

CNCStudent
04-08-2015, 11:05 PM
hmmmm... well I've just read one of the other posts on building (more correctly assembling) PSU's / components and was intrigued by the use of an NTC thermistor for inrush current protection.... (time to do some reading and learning) (http://powerelectronics.com/community/how-do-you-choose-right-type-ntc-thermistor-limit-inrush-current-capacitive-applications)..... [never took it to be much of an issue really .... we just normally just change the type of breaker (D type) to alter the characteristic curve to stop nuisance trips....]

Robin Hewitt
05-08-2015, 11:10 AM
transformer rated at 50v ac, once rectified will provide ~70v DC (with no load), is this not a little high given the geckodrive guidance is 32 x SQRT (L), where L is in mH.... [32xSQRT(3.2) = 57v]

This is the problem I referred to. OTOH I didn't quite get that this thing was designed for an AC input, so maybe I should have kept my wug shut :suspicion:

The problem with rectifying and smoothing a transformer is that you get peak volts at no load and a square wave as you start pulling power. If you try to smooth it the capacitors quickly become bucket sized and hideously expensive.

A switchmode PSU uses the bottom end of the rectified wave form where the low voltage gaps are so much narrower and you can keep the power flow using inductors.

CNCStudent
05-08-2015, 11:40 AM
Any comments are much appreciated, here to learn, just not been impressed by the occasional squeal from my switch mode PSU (given the drives and motors are near silent.... Drives do have fans on them, but no unusual noises, been brilliant drives since recommended on here), did consider the use of a TDK lambda 48v 1000w (FPS1000) PSU as they could be got on fleebay for not a lot and there is also no minimum load requirement, but kind of thought a smooth ac rectified power supply is probably better than a switched mode PSU.

Neale
05-08-2015, 01:01 PM
I wouldn't get too hung up on ripple, within reason. Apart from anything else, Leadshine recommend transformer/rectifier/capacitor supplies rather than switchmode. Ripple isn't too good if you are talking audio amplifiers, but motors aren't that fussy. We are already over-driving the motors from a volts point of view and relying on the drivers to "lose" excess volts. The drivers will have internal regulation (must have, given their wide working voltages) so the internal clever electronics are fine. They have current limiting on the motor drive side and while I don't know what kind of frequency they use for the PWM, I doubt if it will worry about a bit of input ripple at a lower frequency. The downside is that linear PSUs are bigger, heavier, and probably cost more than a cheap Chinese SMPS. But the "soft" current limiting might cause fewer problems with peak pulse loads?

Robin Hewitt
05-08-2015, 08:07 PM
I don't think the problem is in the motors so much as the breakdown voltage of your H bridges :beer: