DC Drive Schematic Concept

Hello everyone,

I am trying to make a 1/2 to 3/4 HP DC drive for my CNC machine. I decided not to buy because I would rather know how to fix it and want to be able to contribute the project to others.

I made a schematic and attached it. It is merely a concept schematic of how I am thinking of doing the drive. I would really appreciate anyone's critiques about the concept itself. I will explain the circuit's 'operation' from left to right.

So, the input comes in from a 120VAC source (isolated) through a 6-8A fuse. The main bulk capacitor(s) will have a very low resistance FET on the return path to negative. This FET will be picked so that it possesses a lot of gate charge, and more will be added to the base. I need to calculate this circuit, but the idea is to use the linear curve of the FET to absorb inrush.

A 15V supply will be derived from the mains through an integrated solution, namely an ST micro Viper IC. I am using an app schematic, so no worries there.

The middle stage is for line detection. A microcontroller will strobe the line every second very briefly by turning on a fet that drives an optoisolator LED (to reduce resistor power dissipation). It looks for the signal on the other end, if the line voltage is there, then it will detect it and start charging the bulk capacitors. Darn, just noticed I am missing some bleed resistors..

The final stage is a simple PWM circuit. The OPTO Gate drive is an integrated 2A optoisolated gate driver IC, I just didnt have the symbol. The mosfet in parallel with the motor acts as a low-resistance diode.

Am I missing anything? I haven't ever done anything this high voltage and potentially lethal. Oh, nobody build this circuit! It is just a concept and is not safe.

Steve

One tiny oversight is the absence of a connection from the bridge rectifier to ground.

I presume the "off" switch is prior to isolation?

Oh yes, thank you! I missed that. Yes, you're correct, I am adding in a manual switch.

I thought of a way to save an opto-isolator, I can use U4 to turn on U2's LED, instead of going through a transistor.

Thanks If no further objections, I will prepare a more calculated schematic.

Steve

Just noting, in case someone would try to point them out:
1) my grounds are messed up, there are two, one isolated, one non-isolated
2) my inrush circuit is a bit off, there are other things I need to do

Don't forget the isolation transformer, don't even test the circuit without one, every wire you touch will be hot!

I plan on testing with an isolation transformer, but in reality, it will be run from the line. I do value my life, so I will definitely be isolated and extremely careful. Thanks for the heads up though! The DC drives on the market are run from the line, since I don't see a 400-800W transformer hiding on them

Steve

Steve, the MOSFET in parallel with the motor will never turn on with the gate connected to the source. The substrate diode will conduct during flyback, but I don't think that's what you want. A diode-connected MOSFET has the source connected to the drain, but it's not very good compared to a diode, and if you swapped source and drain (leaving the gate as is) to do that here, your substrate diode would conduct when you turned on the PWM MOSFET. That would be a bad thing.

Hi Ron,

Thank you for your input, but I fairly certain the MOSFET is connected properly. My basis for this is the following circuit:



and the corresponding explanation

There is a second MOSFET connected across the motor: MOSFETs act like diodes for reverse current, and this is reverse current through the MOSFET, so it conducts. You can use a MOSFET like this (short its gate to its source) or you can use a power diode. However a not so commonly understood fact about MOSFETs is that, when they are turned on, they conduct current in either direction. A conducting MOSFET is resistive to current in either direction and a conducting power MOSFET actually drops less voltage than a forward biased diode so the MOSFET needs less heatsinking and wastes less battery power.

Click to expand...

I am not an expert with FETs, so I could be sadly mistaken. I really wouldn't want to see such a short occur, so to err on the side of caution is what I will continue to do.

Steve

N-channel (enhancement mode, which is your common power MOSFET) MOSFETs only conduct when the gate is more positive than the source by several volts. Yours always has zero volts gate-to-source. As I said previously, the current path you show is through the substrate diode, not through the channel. MOSFETs do make great synchronous rectifiers, but this requires that you apply the proper gate drive at the right time. The quote you posted is correct, but you have not applied the transistor in a manner that will do that.
I'll have a look at 4QD-TEC and see if I can figure out what they meant.
EDIT: I need a link.

Oh well, maybe I will just use a 10A 600V Schottky instead
Better safe than startled !

Steve

You can use a half-bridge, with a p-channel as your (synchronous) flyback "diode", but you have to build in a little dead time in the gate drives to prevent beaucoup shoot-through current. Why do you want to use a MOSFET instead of a diode? Are you looking for more efficiency?

The price is adding up
I will have about 30-40$ in parts to buy, and I could have bought one off of ebay for 50$. I guess it is good to know that I can repair it if it fails. Also, it's part of a larger PCB with all the servo stuff.

Steve

Hi all

I have finished a more professional schematic, and hopefully I can get an opinion or two on the design.

From left to right,

The 120VAC input will be through a chassis mounted switch and fuse/fuse holder. Then, the standard issue 0.22uF 250V capacitor on the input for noise reduction, etc. A beefy 600V 15A bridge, which is bolted to a large heatsink along with two MOSFETs and the flyback diode.

The TVS is rated to turn on fully at 246V and is rated for 1500W. The inrush circuit uses two resistors to charge the capacitor and is timed so that the MOSFET bypasses these resistors. The Mosfet is rated at 250V 60mOhms 40A at 95nC gate charge. Realistically, I do not expect must past 3.5A under steady state, which will give me ~3/4W of dissipation.

The offline non-isolated switcher is used to derive the 16V needed for the opto-gate driver. It can produce about 100mA continuously, which should be enough energy.

Here are the calcs for power requirements:
I want to run 20KHz PWM, which means switching off and on at 20KHz. C = Q/V, = 95nC/16V C ~= 0.6nF, E = 1/2*C*V^2 = 1/2 * 6nF * 16*16 = 0.000000768J , so 40000 times in a second it must charge and discharge, 40000*0.000000768J = 0.03072J/S =~ 30mW, So around 2mA would be needed. Clearly, I am overkilling it. Although, I did ignore resistive losses and inductive energies. Maybe I should have used a zener regulator??

The optogate will provide about 2A of current, so it will be able to turn on the gate in about 50nS, which hopefully won't cause too much losses. Relatively, it is small compared to the 50uS PWM perioud. The flyback diode is a heatsinked 15A 600V schottky

My biggest questions are:

Do I need some sort of input filter other than the capacitor?
Should I be using higher than a 250Vds MOSFET? The reason for using this MOSFET was low conduction losses and I couldn't afford much more than 95nC because my gate drive is fairly weak.
Do I need an RC snubber?

Thanks guys!!!

Steve

P.S.- Hopefully this thing works, so I can share the work with others.

ALSO< PLEASE DO NOT BUILD THIS !! IT IS UNTESTED AND POSES RISKS TO THE INEXPERIENCED!

Steve, where did you get the schematic for your 16V switcher? I don't see anything like that in the datasheet. Did you find an app note, or did you design it yourself, or

Hi Ron,

I found it in AN1357, you can find it here:

http://www.datasheetcatalog.org/datasheet/SGSThomsonMicroelectronics/mXyuxvu.pdf

I will follow their guidelines on making it. They didn't spec the inductors, which is kind of concerning. I will overkill the current handling and lower the resistance of them.

Steve

Note: I'm going to bump up the input capacitors C7 and C8, even though I probably don't need to. The supply doesn't really need to be clean, since it is merely driving the gate and it's threshold is 4V.

Steve

Ron H said:

You can use a half-bridge, with a p-channel as your (synchronous) flyback "diode", but you have to build in a little dead time in the gate drives to prevent beaucoup shoot-through current. Why do you want to use a MOSFET instead of a diode? Are you looking for more efficiency?

Click to expand...

I didn't see this before because I posted at the exact time you did. Sorry for not addressing it Ron

Do you think the half-bridge is a better option? This would be a 'synchronous' type drive, instead of having the steady losses through the flyback diode?

I would like to keep the simplicity of the single FET. I have had issues in the past with high-side drivers and really want to avoid their use. I can generate PWM with deadband from my micro though, which would be kind of nice.

I am really after simplicity and reliability, losing a few watts on a ~600W motor is tolerable. I want people to be able to build this PCB as a 3-axis controller that can be interfaced with Mach3 Mill (hobby cnc controller) and integrates a universal motor / DC motor controller for the spindle.

Steve

scubasteve_911 said:

I didn't see this before because I posted at the exact time you did. Sorry for not addressing it Ron

Do you think the half-bridge is a better option? This would be a 'synchronous' type drive, instead of having the steady losses through the flyback diode?

I would like to keep the simplicity of the single FET. I have had issues in the past with high-side drivers and really want to avoid their use. I can generate PWM with deadband from my micro though, which would be kind of nice.

I am really after simplicity and reliability, losing a few watts on a ~600W motor is tolerable. I want people to be able to build this PCB as a 3-axis controller that can be interfaced with Mach3 Mill (hobby cnc controller) and integrates a universal motor / DC motor controller as a spindle.

Steve

Click to expand...

My personal preference is the diode. As you said, the power wasted as a percentage of the total is small. KISS seems to be a good principle here.

With a 3.5A load, a simulation shows that 60mΩ (the surge limiting MOSFET) in series with the 2200uF filter cap will dissipate 4.5W. Peak repetitive capacitor (and MOSFET) current is about 32A. I think you were estimating 3/4W.

Ron,

I clearly do not really understand what is happening here.

Intuitively, I thought the surge limiting MOSFET Q1 would see an average current of 3.5A. Since it is always on, wouldn't it see an average current current of 3.5A and have a 60milliohm resistance? I just used I^2*R to estimate the conduction losses, at about 3/4W.

I cannot the motor causing the MOSFET's Vds to a point where the on-resistance would change much. Is there some reasoning that I am missing here? I thought these estimates would be fairly solid.

Steve

thank you very much for your valuable time Ron, I really appreciate it! I'm still young and do not see things as clearly as one with as much experience as yourself.