What's new
What's new

PWM Drive for DC Motor

Kezorm

Plastic
Joined
Mar 15, 2016
Probably still a little premature to share, but I'm excited to have been getting some time in on the lathe restoration lately. Project is 1943 10EE that I tore apart almost 15 years ago and am just now getting back to operational. Still lots of paint stripping / repaint in the future (tailstock and all the covers). But, at this point, I don't care what it all looks like. I just want a functioning lathe.

Anyway, part of the restoration is a drive conversion to modern PWM drive of my own design. Developed it years ago for a 3-phase PMAC motor but put the hooks into the design such that it would be able to drive the brushed DC motor in the 10EE when time came. That time is now. Hopefully finish off the power wiring this weekend and can get to work on software and controls.

Input power is 240VAC single phase. Off the shelf 250VDC, 3kW power supply feeds the inverter. 2 legs of the 3-phase drive make up a full h-bridge to drive armature in either direction. 3rd leg feeds the field. When all done, old-school DC motor performance capable of running full-tilt directly off of single phase power :)

ACtC-3e_U5eS30gIJ8XR70uX2Ak5JUP1EP5gIn-OiSby54xGJV2VvAbSCfy_BqkcEIBBiSkPKfNW96VukM5tLs0kA  ZxVC6K8wr8ZlUC7_FHDwGF1HDpFP6qU6q3iOHCQaK62EbWgFSo  04vmHUvbNDl6RuvA=w1280-h960-no

ACtC-3dwveIKT80fEvkhlWCpn6uMyesgr9kywYSeEM3UfD3eZptFFu4  xgfIDHCYDSvh0EFOTHsoTiJ8T5nx-rh9C5hueKHgVtYX5Mees2CT07lVQY7PJRYla_V_6AKTlaSlQCF  s1D0_C8WWOyf2kMqmdke8=w1280-h960-no

ACtC-3e3Oz1YeW28Ms18IWDWqjXTYCVieGwOoKOUgpsYFKsi7YRvP9L  iQ5jn5vqRWS-OJ6QsklN5WeKNsptHO0z9pUtOW4Vi54jXut4Uf_oBJy1P15Kak  1cOjwgZhAqX5fW1k0wG8LISYNHAev0GnMszcNk=w960-h1280-no

 
Last edited:
Cool. I did a PWM drive design on my day job about 25 years ago, 165V 30A, 3 phase brushless DC. Inductive kickback and self generation were a problem.
I would love to see your schematic.

CarlBoyd

Nothing too exciting in the inverter. This is personal project started about 5 years back for PMAC motor. I've done several other motor drives, larger and smaller, over past 15+ years as part of day job. Nothing out there exactly what I needed, so figure I'd spin my own. Also knew I'd want something for the 10EE eventually as I scrapped the motor-generator gear 15 years ago but knew that I wanted to keep DC brushed motor. The 3HP large frame motor is pretty killer in many regards. Maybe not the most efficient thing in the world, but unbelievably smooth with gobs of inertia.

The inverter is 100% design by overkill. No intent to make it a sell-able product. Plan has always been to make everything open source hardware / software, but I'm still hesitant about liability aspects. As you know, 250VDC is nothing to mess around with. I'd be glad to help individuals pursue a build based on the design, but I have absolutely zero interest in everything that would go into trying to produce and sell it.

Anyway, here's the inverter schematic (Schematic - Monarch Motor Power). USE AT YOUR OWN RISK. Everything design by overkill. No attempt whatsoever to optimize cost (inverter is probably around $150 BOM cost at QTY 1). Use off the shelf, oversized IGBT module. Hall effect current sense to make for easy sampling and isolation. Current sense on all 3 legs so that I can use in this Brushed DC application (2 legs for armature and 1 for field). Off the shelf isolated regulators to generate the low voltage logic supply. All the controls are completely isolated electrically from the motor bus voltage. This makes it easy to hook up to the microcontroller for programming, debugging, etc. as the necessary isolation is already baked into the design. Logic board is separate, connected through P1 in middle of page.

TB1 in upper left is DC supply input. TB2 in lower right is 3 phase (and ground) motor output. Everything rated to handle 400VDC bus. Current sense can measure +-40A. I expect the drive would be thermally good for up to about about 5kW. Note that this is 2 quadrant drive, not 4. But, should be all that's needed for 10EE service. (although, theoretically, 4 quadrant would be possible if the supply had ability to sink current)
 
Carl, this looks really interesting. Could you provide the native KiCad project file?

I'm not quite ready to release everything into the wild yet, couple bugs in the current revision that I haven't gotten to fixing yet. But, I would be glad to share individually and help someone pursue their own build if desired. I have several more bare PCBs of each of the boards would be glad to part with for few bucks (mainly just cover shipping). As mentioned, a couple mistakes so they need a little rework, but not too bad.

Do you have account at GitLab? Not ready to make totally public yet, but can give access to individuals. Or, PM me an email or something and I can zip up files to send.

Greg
 
Some renders and actual pictures of the inverter. Size of power board is 4.5" x 6.75". IGBT module on back side of board mounted to aluminum heatsink. Logic board is really just main microcontroller (Microchip SAME51) and a bunch of expansion ports for field I/O. I have two I/O boards designed so far. Both 24VDC industrial digital input boards (EN 61131-2 type 1 inputs). One board is 5 channel digital input direct to microcontroller + USB serial port. This is intended for higher speed inputs like quadrature encoder. Also a STOP input that can be routed direct to the PWM peripheral as hard fault (i.e. hardware shutdown all PWM). 2nd board is 8-channel digital input, interface to microcontroller through SPI bus. Logic board is completely isolated from power. I/O boards all have functional isolation from the logic board.

Greg

ACtC-3elJfwJTaTmRO73gTVoaOFaA5EvAxNtr0Skt0rpo9VK4T2S6FdTI6XkforzUzzN1yB2vFVZvZYEDXCuxBg50XJ5GEjLtE9_vcaEGgqV7huaufmzCzAN6MNa_2UYk9Pc7UuWjUYFGW-FpQDCdXWaGRI=w1512-h1736-no


ACtC-3d3A2qsbkgAjy8pko-iKRCOGqJao4QhtpwCnRUgOA_WPI7Esn1MflfHQH_Tn2ZmY12GRxH-Z5CjJmkg5JF5NbRzcwOSl-K_cg0m76WXP-w_xwMFibT2ZRvsjL_54nA5oej7wVzSDYLrZ2l2Wq1ILos=w1742-h1618-no


ACtC-3cN0OhY_UYIghSrpIEt25MZICZWuWYjpsPcVOC1k3cl-BFyr0NWz5gHVdegIq9dpXsfUJsSoPRC3Ygj48tkAWLCznX5iT3fTf-c81t0uJu8b55oq6cywT3_eiBI7-DtY-F71LatiJ81vvqzgcQnDD8=w960-h1280-no


ACtC-3fdbizFl4P8nMw0pIC1K5v46Ilp_Lr96ngdB4rbUjVykIRAG3XeC4g2qM9l1Vn7Rzy3ughxSopMdwdfHkGDs4UloBVqj2mnDADnECMp-m7N54OVD__wpFSFDsdGp3WzWCyiyC13R1KpnRMG5V2kg3Q=w960-h1280-no


ACtC-3cKEyIyp-YGOzO8tYNsMmU6kNB4VO4fzc9wIf28ChyGTdKCbb6Fd4iYUMSUtQ4uioIkzZiIs1_w1HWumV_LKfQKWQRM_uHJpizZSNtvV4wh7joeZHNULRuiXFP8Iy3p5qUTEc4g3u2V7vEWqRXHq90=w960-h1280-no
 
Did you paint it? What paint & color if so?

Yes, repainted with POR15 "Hardnose" paint (think it's called "2K URETHANE" now). Pretty sure light gray on the base. I honestly forget what state the top half is in - was done over 10 years ago. I thought I just just gotten it to primer, but maybe I put on the dark gray top coat. No idea why I went different color. Think maybe I was planning to put another coat of the dark gray on base to match the top. Sad I can't actually remember. Regardless, good enough for now. The covers and tail housing are still as I bought the lathe, coated in gods know how many layers mostly dark blue. Some day I'll maybe get around to cosmetics, but for now I just want a working lathe, patchwork color scheme be damned.
 
Google-everything is hard-blocked here. No fle for me!

No google?! That's hardcore, man. I won't claim to be fan of the Google either, but damn. Can you see pictures? They are linked from google... (I assume everyone else can see pictures and I don't have some stupid mistake with permissions?)

Will send schematic by email. Let me know if want more details. As mentioned previously, not quite ready to completely release everything to the wild. But, glad to go deeper on individual basis for now.
 
I can see that this is not your first rodeo, and you either have first-class soldering skills or contracted out the assembly. Well done.
 
I can see that this is not your first rodeo, and you either have first-class soldering skills or contracted out the assembly. Well done.

PCB assembly is all hand done in home lab. SMT down to 0402 is easy once you learn the trick. Paste stencils are cheap (<$10). Squeegee on paste with old credit card, hand place components with tweezers (good stereo scope is big help, but not required), pop in converted toaster oven, done.

Greg
 
So OK . you are running THIS PWM creature off 3-Phase input, say 240 VAC leg-to-leg?
This PWM creature running off of 240VAC single phase. DC power comes from off the shelf 250VDC, 3kW supply - https://www.meanwell.com/Upload/PDF/CSP-3000/CSP-3000-SPEC.PDF . Should be just enough for 3HP mechanical. Doubt I'll ever take it anywhere near 3HP in my work. But, if need be, can always parallel another 3kW supply. Or, I can swap in a 380VDC, 5kW supply that I built years back (also 240VAC single phase input).

It will be interesitng to see how well this performs on the 'scope, "real world" use, and how smoothly.
At 8kHz switching frequency, the motor won't know it different from straight DC. Not worried about inverter duty motor. Plenty of thermal headroom in the design, so no need to go excessive on switching frequency / edge rates.

Ignoring time, all in BOM cost is easily under $1000 (including DC supply, filters, inverter, breakers, wire, etc.)
 
That includes the DC PSU? Not bad. But how well does it brake?

Is 2-quadrant drive. Well, technically it's not the inverter that makes it "only" 2 quadrant. The DC power supply is the limitation, can't put it back onto the line. So, will need braking resistor to get better than coast to stop.


None beyond my own use. This is fun exercise for me. No desire to manufacture / sell / support. As you say, no market. I'll be glad to help others investigate the design, but that's as far as I go. I'll gladly share PCB design files, so even the PCB could be considered "off-the-shelf" these days. For example, I paid $72 for QTY 10 of the power board ($7.20 each). Delivery in <2 weeks from China. Gone are the days of photo resist and ferric chloride. It's gotten so easy that I rarely even breadboard anymore. Easier and cheap enough to just whip up quick proto PCB.
 
Gone are the days of photo resist and ferric chloride. It's gotten so easy that I rarely even breadboard anymore. Easier and cheap enough to just whip up quick proto PCB.

Yes, for many years now. I never missed etching or wire wrap. Where I retired from had a world-class PCB shop that charged projects world-class prices. They died gradually, then suddenly. Jobs started going to Shop Overflow, then they even gave up on that and started letting us send things out directly over the InterWeb, and they were bled dry. Our huge world-class machine shops suffered the same fate. Peace is hell on government labs. The Union didn't help, either. The rows of 10EE's, CNC's, EDM's and all the other machines up through huge oil patch-size and vertical-axis lathes went to Reclamation. I hope they went to good homes, and not scrap.
 
Just ran through some quick testing with scope on the field. Everything looks great so far. Little abrupt on the rising edge, will probably throw a bit more resistance on the gate drive to soften it. Plenty of thermal overhead in the design to allow for some inefficiency in the switching.
pic_9_13.jpg
pic_9_7.jpg
pic_9_8.jpg
 
It's alive!

Still lots to do but, for the first time in at least 15 years, it's capable of making chips. First thing I need is actual fwd/rev/speed controls. Right now everything goes through serial terminal on computer (cumbersome not to mention dangerous if I screw up a command). Next, basic IR compensation. Next...
 








 
Back
Top