[contactirfu]

I seriously thought that there is a need to have a separate thread to discuss the RPD, I am planning to have threads linked up here where the RPD has been discussed in the previous UHU thread.

[contactirfu]

The PCB copper thickness is not mentioned in the PCB files, so use at your own risk.

Henrik's schematic supposedly works well.

[NC Cams]

A mor robust circuit can be had by repacing U1 with an LM2940 of the appropriate voltage. 7812's are rather primitive and do not hat the protection already built in like the LM2940 does.

You might want to put some small caps on the inputs to your op amp, at least on the Vref side, as this will make the cut in more precise and repetitive. 0.1 or 0.01 would do.

If you use the 2940, you can get rid of the R1 and D1.

IF the bleeder + has the potential for a lot of voltage ripple, Replace R1 with a 1N4002 and delete D1 completely. IF you do this, reconsider C3 thru C6. Double the app note recommended input big cap at least, I"d use 500uf.

With this, you can also increase the regulated voltage cap size as well. We tend to use a 500 input, 250 output caps, or 100 minimum, with 2940's for very good voltage stability.

[contactirfu]

Thank you NC cams, I am sort of not an electronics person - so I am still learning thru failures partly bcos I fail to understand others failures, when u say 500, 250 and 100 these are uF of the caps right?

RGDS
Irfan

[contactirfu]

NC Cams , I chkd the datasheet for the LM2940, how can we get down the voltage of the motors which is 110V to that required by the LM2940 - are there different variants of LM2940? can youshed more light on this?

[NC Cams]

Most CNC machines have a 24VDC control circuit on them someplace. Use that.

You don't really want to use the servo voltage to power a control circuit. Too many voltage dumps, fluctuations and other stuff that simply causes havoc with control circuits. The 2940 should work fine with 18-24 input to get 12vdc output.

If you HAVE to use 110vdc input, check out Nationals high voltage regulators for some ideas. If it were me, I'd find and use the 24vdc that is in the machine SOMEPLACE!!!

Finally, yes, the cap numbers are in uF.

[contactirfu]

thanks SIR, you are right.

[Mariss Freimanis]

My 2 cents:

The circuit should work OK but its drawback is it's set to a specific voltage. What if the supply voltage changes or someone mis-sets the trip voltage? The consequence is the dump resistor will dissipate large amounts of power continuously.:-(

A better solution is to sense power supply current direction. As long as current flows from supply to the drive, the dump is 'off'. When current flows from drive to power supply, the dump is 'on'. This makes the dump circuit independent of supply voltage and guarantees the dump resistor cannot conduct continuously (the drive can only source current until the stored mechanical kinetic energy is exhausted). Also there is no trimpot to set (or mis-set).

Mariss

[contactirfu]

Mariss, thanks for the inputs.

you mean that sensing the reverse flow is always better, is there any circuit which I can build/buy for myself? do you know anything which can operate above 100V

RGDS'
Irfan

[Mariss Freimanis]

Well, 'better' to me means something that is simple to build, solves several problems simultaneously and cannot be misadjusted. The attached dump1.pdf and dump2.pdf circuits dump returned energy and secondarily, rapidly discharge the power supply when AC is switched off. No 'bleed resistor' is needed; the supply discharges in less than a second. I threw in e-stop dump.pdf just for the heck of it.

Mariss

[NC Cams]

My 70's vintage Fanuc 5T does essentially what Marris recommends.

They sense the drive logic signal which ia based on +/-10vdc - I suspect they use a window comparator circuit as this is a simple, effective way to sense the dead band of the servo drive. They then use this to trigger a relay that serparates the drive from the servo and simultaneously dumps the drive onto a regen power absorbing resistor.

The relay is a big clunking NO/NC that handles the swiching of the regen/motor/drive circuitry as needed.

Dumping the power from the motors is the easy part - how you prevent the power from frying things is the hard part. The dumping of 4 servo's across one dissipation resistor doesn't seem like a valid approach to me. I was hoping that other more qualified souls would have chimed in on that.

Besides, the curcuit in this instance works fine if the motors turn CCW - what happens if the motor is turning CW and the power to be dissipated is of the opposite polarity? I venture smoked fets in the dump circuit.

Take a look at the Fanuc method of power dissipation from say a 5M or 5T. No sense reinventing the wheel.

Trying to dissipate the power once it is on the drive buss is a bit late. It is better to isolate or keep the power to be regenerated from getting to the drive BEFORE you have to deal with it. THis is why Fanuc sensed the drive logic as opposed to the drive bus voltage for their regen circuit.

[contactirfu]

this sounds interesting - but what would be the exact solution we can use is what we all HP UHU user are looking at.

I am so illiterate in these things that I am still trying to make sense of what is being said.

RGDS
Irfan

[rokag3]

i build this circuit below made by Mariss Freimanis of gecko
http://www.hv4all.com/cnc/Power%2520...%2520Bleed.pdf
it work fine to discharge the capacitor (i use 2 bulb 220v 100W in // r=40ohms per bulb =20 ohms)

[Mariss Freimanis]

While on the subject: A brush-type servo doesn't need returned energy protection because it cannot generate any in normal operation. The motor 'back EMF' is proportional to speed, speed is limited by the supply voltage, therefore all voltages are equal to or less than supply voltage. No 'returned energy'.

Two exceptions:

1) There is energy stored in the motor's inductance. This is a minuscule amount compared to stored mechanical energy and can be safely neglected.

2) During an e-stop you do want to stop the motor quickly instead of leisurely coasting to a stop. The original circuit doesn't address this while all three of the circuits I posted do. The 'dump' circuits dynamically brake the motors to a stop when AC is removed to the power supply. The motor becomes a generator working into a large load (dump resistor) which quickly extracts stored mechanical energy from the motor and mechanism.

Mariss

[Mariss Freimanis]

rokag3,

Remember a 220VAC 100W light bulb has a 'hot' resistance of 20 Ohms. The 'cold' resistance is much lower, typically 2 Ohms. Use an Ohmmeter on this light bulb to verify my comment.

Mariss

[Mariss Freimanis]

NC Cams,

It is a perfectly valid method. The currents and voltages look impressive but the amount of energy involved is relatively small meaning the dissipation event lasts a short time. As an example, a 1,000 lb mass moving at 500 IPM contains only 10 Joules of kinetic energy. A Joule is 1 Watt for 1 second; decelerating this mass to a stop in 1/10 of a second means the dump resistor dissipates 100 Watts for 1/10 of a second.

The dump resistor can be as small as 5W for this example provided it's a wire-wound resistor. Wire-wound resistors have the highest pulse power dissipation rating, typically 100:1. This means a 5W wire-wound resistor can dissipate 500W peak power for a short period of time without any damage.

Mariss

[H.O]

Hi guys,
Since I'm the one who originally designed and published the circuit that Irfan has attached to the first message in this post I thought I'd explain a bit why I did it the way I did....

Before designing the circuit I looked at both of Mariss' circuits that he attached to his post (they've both been available on Geckodrive's Yahoo-group for quite some time).

Dump1 was discarded because I couldn't source a PNP transistor rated at 150-170V and ~30A. Dump2 was also discarded because I'm using a three-phase rectifier and I honestly couldn't figure out if that circuit would work with a three phase rectifier. (I confess, I'm embaressed). I did post a question on the Geckodrive Yahoogroup about it but it was/is unanswered so I set out to design my own.

Also, on the schematic you see two connections marked Bleeder+ and Bleeder-. In my own system these are connected to a 7.5ohm resistor via the normally closed contacts of the main contactor. When entering E-stop the contactor deactivates and connects the resistor across the powersupply capacitors, discharges them and "breaks" the motors. Agreed, it requires a second resistor....


Mariss, I'm a bit confused that you say a revese energy dump isn't needed under normal conditions. When testing my system I set the dump-circuit to activate at 15V above nominal powersupply voltage (nominal is 130VDC) and it clearly activates during decceleration from high speed. There's even a video on Youtube (not mine but he's using "my" dump circuit) where you can see the difference with and without the circuit. I'll see if I can find it but I think I must have misunderstood you or something.

/Henrik.

[Mariss Freimanis]

Henrik,

First let me say you have a well worked out circuit so my comments were not intended as any kind of negative criticism.

Somehow I missed your post; I would have responded otherwise.

Dump1 circuit: You can replace the TIP147 with a small-signal PNP like an MPSA92 which has a Vco of 350VDC. Install a collector resistor sufficient to limit collector current to 1mA or 2mA at 170VDC. Have the end of this resistor drive the gate of a suitable n-channel MOSFET whose drain circuit contains the dump resistor. MPSA92 collector -> 100K -> MOSFET gate, MOSFET gate -> 10K -> GND. Place a safety 12V zener gate -> cathode, anode -> GND.

Dump2 circuit: The main advantage over dump1 is the wasteful rectifier in dump1 is eliminated. It will work with 3-phase as well as with single-phase.

Returned energy: Never post until the brain is fully saturated with caffeine; it wasn't and I wasn't thinking. The theoretical maximum returned voltage is twice the supply voltage, i.e. motor back EMF equal to supply voltage, '-' side of motor to 'on' top bridge MOSFET.

I have attached a diode decoupled supply voltage scope trace going to a G320. Supply voltage is 48VDC, NEMA34 motor is decelerating from 3,000 RPM to zero in 100 milliseconds. Note the voltage going to near 80VDC, 30 volts higher than supply.

Mariss

[H.O]

Hi,

First let me say you have a well worked out circuit so my comments were not intended as any kind of negative criticism.

Not taken as such - just explaining the reasons the circuit is what it is. It works and it has served me well but there's always room for improvement... Good to know that the Dump2 circuit does indeed work with a three-phase rectifier, thanks!

/Henrik.

[Mariss Freimanis]

Let me hasten to add that another 1N4004 going to the bridge must be added in the circuit. A single-phase full bridge rectifier has 4 diodes, a 3-phase bridge has 6 of course.

Mariss