Regenerative power dump circuit

[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