[H.O]

Thanks Irfan,

I was really hoping to get more comments, feedback etc about this to help me and everybody else to understand.

I have no reason at all to doubt Kreutz, after all, he designed the thing, but I'm getting quite concerned about what the drive actually is capable of. IIRC the drive will be rated at around 160V, 20A when infact you won't be able to drive anything near that without crippling the motor performance with HUGE resistors in series with the armature.

For example, Kreutz told me that the BEMF+0.87*PowerSupplyVoltage should be kept below 200V. (I don't know why but that's what he said). In that case, how in the world would you be able to run a 130V motor? A 130V motor obviously need 130V to develop full power. To get 130V at the motor (disregarding the I*R drop in the armature) the powersupply voltage needs to be 150V so then we have 150*0.87+130=260 which is WAY over 200....I don't get it...

I know you and Paul aren't doing this for a profit but (IMHO) if this is going to work for you you absolutely have to get the facts around this straight and publish it somewhere. Otherwise people like me will buy the drive, connect a powersupply and motor well within the rating of the drive and even so taking a rather big risk of it blowing up. I know my particular motors are borderline on size anyway but concidering the HomeShopCNC motors are WELL within 160V, 20A but doing the math on it shows that is about what you can safely run.

I'm really sorry for the way this post sounds. It's not meant as criticism towards you, Paul or Kreutz I'm just trying to get this straight before people buys it and gets dissapointed - as have happend before.

Thanks for listening!

/Henrik.

[contactirfu]

..............

[kreutz]

Hello,

I am preparing a pdf document with explanation to the issue and its cause. The 200 V limit is a derated working limit for the Mosfets, and also due to the 470 uF/200V capacitor used on the power supply input. If you use regenerative energy dumping it is going to be less critical.

Regards,

Kreutz

[H.O]

Kreutz,
I think many people will find that very usefull - I know I will, thank you!

Irfan,
I'm not quite following you on that last message... I can only imagine the amount of time and effort you and Paul, not to mention Kreutz have been putting into this - I think I can speak for everybody here when I say that we appreciate it! I'm just trying to figure this thing out so that I don't get dissapointed later on....as always the devil IS in the details.

/Henrik.

[kreutz]

................ I'm just trying to figure this thing out so that I don't get dissapointed later on....as always the devil IS in the details.

/Henrik.

I appreciate the fact that you brought the subject to the thread. The more everybody knows about motor control, the better. Knowledge helps when you need to make decision, and you know what you are looking for.

Thanks,

Kreutz.

[kreutz]

Here is a preliminary release of the promised document.
I really haven't had time to review it because of the limited time available. I will correct it if mistakes are found or more clarification is needed in a certain area. I should note here that the worst case scenario of power stage Mosfet destruction during sudden direction reversal at high speeds, does not present itself often in real life CNC controlled servo due to CNC software controlling acceleration and deceleration before a direction change. It is more prone to happen during motor PID parameter tuning.

Thanks,

Kreutz.

[H.O]

Kreutz,
Thank you for taking the time to write that up! It's getting clearer and clearer.

Just a small note:
On page 3, in the forward example, you say that the upper left transistor (Q1) is ON and the lower right transistor (Q4) is PWM'in and the current thru the motor is flowing from right to left. Shouldn't that be from left to right?

Thanks!
/Henrik.

[kreutz]

Kreutz,
Thank you for taking the time to write that up! It's getting clearer and clearer.

Just a small note:
On page 3, in the forward example, you say that the upper left transistor (Q1) is ON and the lower right transistor (Q4) is PWM'in and the current thru the motor is flowing from right to left. Shouldn't that be from left to right?

Thanks!
/Henrik.

Thanks a lot, the document has been corrected.

Regards,

Kreutz.

[LtcFisher]

Hello everyone, I'm new here and have been reading about the problem of Breaking for DC motors and blowing the mossfet's. What I did, if it is worth it was to add to relays into the circuit ( 1-Forward, 1-Reverse ) during normal operation they will supply the motor with power, when not in use they dump the regenerated current being produced by the motor during coasting to a grid bank of 3-resistors. Hope it helps. I might be off track on what your trying to do.

[kreutz]

Obsolete parts notification

I've just received a note from Digikey telling me that the component HER105-T (part number: HER105DICT-ND) has been declared obsolete, but still available in limited quantities. Substitute is Digikey part number: UF1004DICT-ND (which is a few cents more expensive).

I will replace the part number on the next revisions of the schematics.

Thanks.

Kreutz.

[kreutz]

Hello everyone, I'm new here and have been reading about the problem of Breaking for DC motors and blowing the mossfet's. What I did, if it is worth it was to add to relays into the circuit ( 1-Forward, 1-Reverse ) during normal operation they will supply the motor with power, when not in use they dump the regenerated current being produced by the motor during coasting to a grid bank of 3-resistors. Hope it helps. I might be off track on what your trying to do.

Hello;

Welcome to the thread.

While yours it is a good solution for normal DC motors, it is more un-reliable to implement on positioning servo motors (which are constantly changing direction under software command like in a CNC machine) due to the high frequency of operation of the relays. Adding an electronic regenerative energy dumping circuit should help because it is faster, there are no relay contacts to replace, and only dumps excess energy when required (above certain voltage level on the capacitors)

Thanks for sharing.

Kreutz.

[ckm]

Hello;

Welcome to the thread.

While yours it is a good solution for normal DC motors, it is more un-reliable to implement on positioning servo motors (which are constantly changing direction under software command like in a CNC machine) due to the high frequency of operation of the relays. Adding an electronic regenerative energy dumping circuit should help because it is faster, there are no relay contacts to replace, and only dumps excess energy when required (above certain voltage level on the capacitors)

The way this is implemented in my Servo Dynamics spindle drive is via a set of power MOSFETs that switch inrush current to a series of resistors.

Actually, the power section of my servo drive is currently broken, with the drive sensing an over-current condition on the power feed, likely due to either a broken rectifier diode or the regen switching not working. It's really hard to bench test either (the power section requires 3-phase AC...), but I've been muddling through fixing it by replacing components that might be broken...

Chris.

[vrsculptor]

I think I finally understand the UHU’s braking resistor. During acceleration the UHU sends electricity to the motor through the H-bridge. During controlled deceleration the UHU does not apply braking but instead uses the motor as a generator to decelerate it. The motor’s internal resistance slows the motor and as long as the controller’s deceleration curve matches the ability of the drive to handle the returned energy and or the capacitor to store it everything is fine.

In the case of extreme deceleration caused by a metal to metal crash the drive should not be hurt. It will fault due to position or over current. Current re-circulation is not an issue as the machine’s structure absorbs all of the excess energy.

The problem is with an E-stop. I have a 500 pound table doing a rapid at 200 IPM and the E-stop is triggered which shuts down all power to the machine. The uncontrolled deceleration pumps a lot more energy into the UHU than it can possibly handle and the FET’s melt and the capacitor explodes.

With my Gecko’s I handle E-stop by shorting the power supply through a resistor/contactor and everything stops pretty quickly. Dropping AC power won’t help the UHU as current will still be passing through the H-bridge. Could you use a DPDT relay to disconnect the motor from the drive and connect it to a braking resistor (150-watt light bulb) on an E-stop without destroying the drive? Would this solve the problem?

[ckm]

I think I finally understand the UHU’s braking resistor. During acceleration the UHU sends electricity to the motor through the H-bridge. During controlled deceleration the UHU does not apply braking but instead uses the motor as a generator to decelerate it.

Yes and no. Every spinning motor acts as a generator to some extent. When you cut the power to a motor, inertia will take over and the motor will spin down naturally. During spin down, it generates electricity and servos are particularly good at generating electricity. This electricity needs to go somewhere, as feeding it back through the controller would cause all kinds of problems. This problem is usually handled by a 'shunt regulator', which switches the path of the motor leads to a set of resistors. The resistors basically heat up and this 'consumes' the electricity. Pretty much any electric motor of any size will use this mechanism, including electric locomotives.

The resistors do slow down the motor a little, but that is not a reliable method. If you really want to slow the motor down, you use a brake on the shaft, which is usually activated when the electricity supply to the motor is stopped. However, suddenly shutting off power to the motor and stopping it with a brake will cause large amounts of back EMF. Think of a fast flowing river, then dropping a dam across it all of a sudden. Large waves and water pilling up would be the result, same with electricity. Most shunt regulators will deal with back EMF as well (it looks the same as motor spin down for all practical purposes...).

If you are worried about a crash, use a motor brake + shunt regulator and have the brake be controlled by presence of electricity to the motor. The shunt regulator should be MOSFET driven or other high-speed switching transistor, not with a relay. The brake can be controlled by a relay.

Some useful links.
http://www.elmomc.com/applications/D...-regulator.htm

http://www.discovercircuits.com/H-Corner/dcbrake.htm

http://pdf.directindustry.com/pdf/ne...614-14286.html

http://www.electroid.com/

HTH,

Chris.

[bigandy]

Good evening all.

I am looking to build a handful (3 plus one spare) UHU servo drives, and I already have the programmed chips from Uli. I was wondering if anyone is currently selling kits of parts and PCBs to build up the drives? I'm going to be using the drives with some Dunker motors rated at 3600rpm@60v and 2amps under no load, and IIRC the stall current is around 19 amps.

Thanks in advance
Andy

[kreutz]

I think I finally understand the UHU’s braking resistor. During acceleration the UHU sends electricity to the motor through the H-bridge. During controlled deceleration the UHU does not apply braking but instead uses the motor as a generator to decelerate it. The motor’s internal resistance slows the motor and as long as the controller’s deceleration curve matches the ability of the drive to handle the returned energy and or the capacitor to store it everything is fine.

In the case of extreme deceleration caused by a metal to metal crash the drive should not be hurt. It will fault due to position or over current. Current re-circulation is not an issue as the machine’s structure absorbs all of the excess energy.

The problem is with an E-stop. I have a 500 pound table doing a rapid at 200 IPM and the E-stop is triggered which shuts down all power to the machine. The uncontrolled deceleration pumps a lot more energy into the UHU than it can possibly handle and the FET’s melt and the capacitor explodes.

With my Gecko’s I handle E-stop by shorting the power supply through a resistor/contactor and everything stops pretty quickly. Dropping AC power won’t help the UHU as current will still be passing through the H-bridge. Could you use a DPDT relay to disconnect the motor from the drive and connect it to a braking resistor (150-watt light bulb) on an E-stop without destroying the drive? Would this solve the problem?

On the E-stop case, if you shut-down the power supply voltage you need to have a regenerative energy dumping circuit (still active) to handle the excess energy stored on the load. I don't have much information about the Gecko drives, but I think they will behave the same as the UHU drive under similar conditions.

So far I haven't seen/heard of any drives, mentioned on the CNCzone, having regenerative current control.

During controlled deceleration the UHU does not apply braking but instead uses the motor as a generator to decelerate it

All the drives using Four-Quadrant PWM will behave the same, all of them return excess energy to the power supply during deceleration, either by using Synchronous rectification or by using the anti-parallel diodes as a path for the regenerative current.

Disconnecting the motor from the drive will destroy the H-bridge faster. You can use a load resistor in parallel (connected using a contactor) to the motor armature to help dissipate regenerative energy (and to act as a brake too). The only problem is the time interval between power supply AC disconnection and load resistor connection (msec), during that interval the current will flow through the drive to the bus capacitor.

Normal braking, even under such circumstances, will not destroy the drive unless the motor internal resistance value is so small that does not limit peak current through the H bridge anti-parallel diodes to a value within the Mosfets' "Safe Operating Area". If damage happens during normal deceleration (without reversal) the drive is working outside its manufacturer's suggested specs.

"Braking Resistor" is a term applied basically to the permanent "series Connected" resistor used to limit peak current when suddenly reversing direction when moving forward at maximum RPMs and under maximum power supply voltage.

Kreutz.

[Xerxes]

However, suddenly shutting off power to the motor and stopping it with a brake will cause large amounts of back EMF. Think of a fast flowing river, then dropping a dam across it all of a sudden. Large waves and water pilling up would be the result, same with electricity.

I think this is not correct. EMF is proportional to rotation speed, so if motor is stopped fast, then also EMF drops to zero. So motor does nothing nasty during sudden mechanical stop.

During crash the real problem is drive that still thinks that motor has lots of EMF and keeps pushing lots of voltage to motor. In reality drive is fighting against winding resistance and zero EMF. Eventually this overloads the power stage and magic smoke escapes. This happens only on "blind" drives that do not have continuous current/torque control which would keep eye on motor current.

[kreutz]

...suddenly shutting off power to the motor and stopping it with a brake will cause large amounts of back EMF...

Nothing bad will happen, zero speed = zero BEMF.

The energy restored to the power supply comes from the inertia of the mechanical load driving the motor axle while no power is being applied from the drive side (that is why it is called re-generative energy, the motor acts as a generator).

By the way, most of the electro-magnetic brakes used on servo motors are there only to avoid the load from driving the motor axle during "stand-by" in a stop position (static brake), they are not used as a dynamic brake, not even intended for that job, and will quickly deteriorate if used to brake a moving rotor. Some drives will have a "brake" command, driving both of the H-bridge's lower (or both upper) Mosfets ON at the same time, while the motor is stopped or rotating under certain low RPM limit, acting as a dynamic brake.

[vrsculptor]

Normal braking, even under such circumstances, will not destroy the drive unless the motor internal resistance value is so small that does not limit peak current through the H bridge anti-parallel diodes to a value within the Mosfets' "Safe Operating Area". If damage happens during normal deceleration (without reversal) the drive is working outside its manufacturer's suggested specs.

Kreutz,
Thanks for the guidance. I'm a software person and not an EE. So far I've been pretty successful building power supplies and wiring Gecko's and encoders without needing any deep understanding. I have not had to get very theoretical. I will have to think on this and do some further investigation to see if I am in trouble or not.

My original servo amps were rated at 20 amps continuous, 160V DC and would like to be able to stay in that ballpark. The servo's are about 3.5 ohm's.

[ckm]

Nothing bad will happen, zero speed = zero BEMF.

The energy restored to the power supply comes from the inertia of the mechanical load driving the motor axle while no power is being applied from the drive side (that is why it is called re-generative energy, the motor acts as a generator).

Wrong terminology, sorry, that's what I meant.

By the way, most of the electro-magnetic brakes used on servo motors are there only to avoid the load from driving the motor axle during "stand-by" in a stop position (static brake), they are not used as a dynamic brake, not even intended for that job, and will quickly deteriorate if used to brake a moving rotor.

Yes, but he was referring to an e-stop. The servo on my 6hp spindle has a brake on it that engages every time the spindle is stopped (e-stop or just turned off). AFAIK, it could be considered a dynamic brake and it's still working fine even though the machine is 20 years old. And most brakes I have seen, even on small servos, are wired such that they engage as soon as the power is shut off, so they do brake somewhat dynamically. Then again, I haven't seen that many brakes (a 1/2 dozen), so I could be wrong.

Chris.