Must i isolate?

[abfer]

I want to design a dc motor controller.(300W, 24V dc motor) I plan to use H-Bridge. What i wonder is, must i isolate control circuit from the power circuit completely? I don't want it to be cost much. If i isolate everything then the circuit will be more complex and will cost much. What do you suggest me? Can i obtain a reliable drive without isolation? If i isolate everything, what can i use for voltage and current measurements which doesn't cost much.

 

[Warpspeed]

The problem will not be the drive, but the motor.

It is unlikely that the insulation within a 24v motor will safely withstand full mains voltage between windings and frame for any length of time. That could be very dangerous.

 

[itsmoked]

No you don't have to isolate that. Just pay close attention to wire routing and circuit layout. Make sure you keep the noisy 24VDC stuff away from your control stuff. Include chokes and filter caps where reasonable.

I'm lost on Warp's response "full mains"?

 

[Warpspeed]

Hehehe, the way it was worded, for some reason I just assumed it was a direct off line transformer less system, otherwise why would there be any need to completely isolate part of it from the "power circuit" ?

Sorry, my misunderstanding.

 

[abfer]

Ok thanks. Motor won't be fed directly from the utility grid.

 

[Warpspeed]

abfer, I agree no need to completely galvanically isolate the power system from the control circuit, but beware of induced electrical noise.

Probably the main thing to avoid is using a common return wire for both dc power to the motor, and as a signal return wire. There should be a "power circuit" and a "control circuit" that have quite independent supply and return systems. These two grounds can be connected together but only at a single point.

The rest is all about good wiring practice, bypassing, and even cable screening if the cable runs are very long.

 

[itsmoked]

Warpspeed is correct miss cabling is a constant problem with these type systems. Watch the grounds and keep the separations. You can trade separation and wire routing for isolation up to a point.

 

[ScottyUK]

Depending on the relative complexity and cost of the control cirucitry to the power stage, isolation has the advantage of limiting the damage caused by failure of the power stage. Of course, the advantage gained must be traded aginst the cost of providing isolation and the likelihood of failure.



----------------------------------

If we learn from our mistakes,
I'm getting a great education!

 

[registigger]

I agree with our colleagues above. Isolation is not mantatory but here are a few more details to take into consideration.

1-If your product is medical or in a wet environment, you might have to isolate anyway to prevent exposing operators or patient to high frequency current when touching the controls.

2-If you use a micro processor, the capacitive coupling with the "grounded" environment could cause problems. Make sure that all the lines are terminated properly.

3-If you use potentiometers and switches, make sure that they can handle the high frequency voltage spikes that will be present between their circuit and operator. Membrane switches could be a problem.

Otherwise it may work well.

 

[itsmoked]

marcosheila brings up many good details.

abfer you have not stated whether this is a product or a one-off.

The point I want to make tho is if this is a system YOU will be installing, and can control the wiring, you can get away with out isolation, barring failure saftey protection as mentioned above. But if this is something installed by others, say this is an off-the-shelf product you will sell to others, ISOLATE IT! Otherwise some systems will work, others won't, and worse some will sorta work! You will find yourself on the phone for the rest of your life talking to the unhappy ;(

 

[abfer]

Thanks to all. Yes itsmoked, that would be a boring situation to face frequent failures as it is thought to be a product. But installation will be made by us. But it will be operated by others. To marcosheila, it won't be a medical product or will operate in a wet operation. Also it won't be in a directly reached place nor have a life critical function. It will be in a closed box. Isolating adds real difficult and cost for me so i want to eliminate it if possible. I'll care to connect different gorunds at one point. Should i insert some inductance(or other things) between grounds? How about inductance value than?

"2-If you use a micro processor, the capacitive coupling with the "grounded" environment could cause problems. Make sure that all the lines are terminated properly"

Can you please explain more detail this or suggest me a source to help me learn this subject.

Actually i made a drive simpler than this circuit without islation with one switch. It doesn't contain sensitive measurements. Measurements affected from the high frequency noise however. As they weren't sensitive i solved them by software. But i didn't faced serious problems or failures in control circuit. It has 3 different grounds interconnecting at one point. I think if i care more to reduce noise i can success.

 

[Warpspeed]

A microprocessor will always be far less tolerant of electrical noise than an analog circuit, and special care needs to be taken.

One very effective way is to fit opto isolators close to the processor, and run individual twisted pair two wire circuits back to the motor. Proportional control or measurement can be via pulse width modulation or voltage to frequency conversion so that the opto isolators operate in digital mode. That is always the best and safest way.

The whole thing need not be isolated totally, but isolate all the long wiring runs at the processor end, that are noise sensitive.