routercnc_MK3

[Jonathan]

Silyavski,
Thank you for that link - very good reading. Based on Jonathan's comments and your selection of the 40,000uF model then based on Neil's suggestion then you might only be getting 10,000uF (10,000uF x 2 /2).

Sounds like it to me.

Any issues with running like that on your machine? If I use my board as-is then I'll only get 15,000uF.

Sizing the capacitors is quite tricky - there are formulas floating about on forums, but prizes for finding one with some derivation or reference, not just a rule-of-thumb. You can easily calculate the voltage ripple on the capacitors for a known current (V=I/(C*f)), but what current do you use and what voltage ripple is acceptable? We can get a reasonable idea of the input current from the motor power ratings, but even then it's a bit of a guess as to what voltage ripple the stepper driver will tolerate. You could look at it in terms of what voltage (and hence current) ripple the capacitors will tolerate, or define some minimum input voltage from the transformer you'll allow and go from there. I'd probably go with the former...then take an educated guess!

For a while I had my machine running 4 motors from a 500VA transformer with 11000uF capacitance and I noticed an improvement in obtainable feedrates when I moved one motor to another (320VA) transformer and left the original transformer and capacitors with just 3 motors. So maybe 15000uF is fine ... bit subjective though as I didn't take any readings to verify.

As for modifying the board to get the full 60,000uF - the tracks are 'painted over' ; either that or it is a multilayer board because apart from the solder points there are no exposed tracks on the top or bottom. I'd probably have to demount the capacitors and run flying leads to them from the output of the rectifier.

That's a shame. Do you have a working CNC router, or is this your first? If you want I can send you the dxf to make these ... it's very simple:




Or we could do some sort of swap.

Mains was 243V AC and each secondary coil was 26.4V AC

That's normal - they are a bit high as the rating is when loaded.

I need to wire the secondarys in series but cannot find any markings on the cables to show which is the start of the winding. I've read that it does not matter how they are connected (unlike parallel which is vital). Can someone confirm this is the case?

Not the case. If you get them the wrong way round, the voltages will cancel out and you'll get approximately zero. I would guess that the order they are written on the label is indicative of their polarity, so try connecting red to orange and measure the open circuit voltage between black and yellow.

[routercnc]

Hi Jonathan,

Thank you for taking time to write that comprehensive reply. I'd done a bit of thinking on the series connections and quickly realised that one way would be ~0V and the other way ~50V AC ! So thank you for confirming that. I'll try connecting as you suggest - its either one way or the other and at least being series it won't matter if I get it wrong.

Had another look at the regulator board and I can now see some tracks but am not confident I could make a safe job of modifying it.




So I've also decided to de-mount the capacitors and build a more conventional system. I've ordered:
Solder sucker (had one years ago but can't find it ~£4)
35mm diameter capacitor clamps (6off ~£1.50 for a pair)
400V 35A rectifier centre bolt mounted (~£3)

Thank you for the kind offer to swap / supply dwgs for the capacitor circuit but I'm going to go for the traditional route of mounting each capacitor in the plastic mounts (inverted) and solder cables between the pins. It looks like there will be about 3mm of pin remaining from the top of the caps which should be enough to a cable get onto. The rectifier will be the traditional 4 spade pin type. All bits will be mounted to a ~4mm aluminium plate for heat dissipation and general sturdiness.

That's good background info on the ripple voltage. I had started to wonder about what an acceptable level would be. More capacitors will be better with lower ripple but I guess gives more in-rush current so more chance of tripping the MCB on start up. At least I can play around with the number of capacitors if I have that problem. But if it just works then I'll leave it and get on with using the machine.

For info this is my 3rd generation of cnc machine, with many minor iterations of development along the way, but it remains my only cnc machine and is currently out of action until the PSU if fixed.

[Jonathan]

Don't you just hate black soldermask :|

If you remove three of the capacitors that are in parallel, then put them back in the other way round, you will have one terminal of all 6 connected, so just one piece of wire will be needed to connect the '+VCC' and '-VCC' terminals to have the capacitors in parallel. If you can break one track on the PCB and add one piece of wire, then you can use the existing diodes.

More capacitors will be better with lower ripple but I guess gives more in-rush current so more chance of tripping the MCB on start up. At least I can play around with the number of capacitors if I have that problem. But if it just works then I'll leave it and get on with using the machine.

Yes, we'll see - you will have a lot of capacitance. I looked at that situation recently with my motor drive - that's got 660uF @ ~340V and with direct connection to the mains the scope showed that the current peaks at about 15-30A depending on the timing. That still doesn't blow a 6.3A slow blow fuse though, as the duration is only a few milliseconds.

[routercnc]

Hi Jonathan,

Thank you again. I had to draw out your idea of reversing the capacitors before I could see how it could work. I considered it for quite a while but in the end was concerned about breaking the tracks and adding flying leads. I've had enough of blue flashes and bangs.

I've managed to remove the capacitors from the board. Looks like there is plenty of remaining pin to solder the cable to:


Capacitor mounting brackets and a new rectifier should be here by the weekend so I will lay it all out on the mounting plate if I get the opportunity.

I've decided for reasons of space (6off 10,000uF caps with mounting brackets takes up a surprising amount of space!) and in-rush current to go with 40,000uF as a starting point.

I'll report back with success or more questions . . .

One more thing, I've added a safety latch relay for the spindle inhibit. This makes sure that if there is an e-stop (which stops everything including the spindle) that when the e-stop is cleared the spindle does not fire up until the start button is deliberately pressed again. Ordered 2 pole DIN relay for this job. Drawing updated:

[routercnc]

Based on a discussion I was having in another thread (but didn't want to hijack) my plan is to remove power directly from the drivers in the event of an e-stop via relays.

But as the steppers a likely to be moving then is this is problem for the drivers? I'm sending an e-stop to the BOB at the same time so the motors should be in the process of stopping.

Would it help if I added a capacitor across the relay coils to slightly delay driver shut down prior to the motors coming to rest - just a few milliseconds? Or am I worrying over nothing?

Does anyone else remove power to the drivers using relays in the event of an e-stop?

[JAZZCNC]

Or am I worrying over nothing?

Does anyone else remove power to the drivers using relays in the event of an e-stop?

Yes your worrying over nothing.!!! . . . If you remove signals so no pulses are still getting thru while caps drain then just turning off the power will be enough and motors will come to stop quick enough.

Edit: If you do want to go a little OTT then have e-stop turn off drive Enables so motors are still under torque and stop quick then remove power second or two later using a timer relay.!!

[routercnc]

Hi Dean

In my layout (see post above) my relays are AFTER the PSU so there are no caps to drain and the drivers loose power immediately. So I either leave it, change to a relay BEFORE the PSU on the mains side which I think is what you are referring to, or add delay / capacitor across the existing driver relays on the DC side to hold them closed for a few milliseconds. Hope this distinction makes sense.