PINNED - Roboteq Controller - developing for powerchairs

[snaker]

@Vitolds: According to the document AN70614, just the emergency switch alone is already overcomplicated with too many components. Is it realistic? Does anyone follow all the document recommends? And it still have not stated about how to use the contactor safety pin yet.

[JMGarage]

@Snaker: You have mixed these two pics. If You mean that diode over fuse on second pic, that is for protect for when fuse blows and regenarate occurs. Then You have still connection to battery back and don't broke controller.

First pic is for DOUT.

Then there is STO feature on newer controllers for emergency stop. It should be safe way stop all. Older controller doesn't have this.

I found CANBed V1 what I am going to use.
https://www.longan-labs.cc/1030008.html

[Burgerman]

I have that diode (actually a solar stud type ones, 100A) to provide a path back on the main fuse as well as the contactor. And TVS diodes across the roboteq in and outputs.

[JMGarage]

Have You some code or link for these diodes? Do You have some schematic of wirings and button arrange?

[Burgerman]

No.
But no need.
Ebay TVS diodes, use on all inputs and outputs on the robiteq to clip spikes from damaging the inputs. Choose the voltage above your working voltage and below the roboteqs max input voltage limit.
For the diodes across the fuse, and across the contactor use any solar type diode 30A or so is enough as it is just a short term safety so they can cope.

Wiring as per roboteq manual.

I also used 2 SSRs - dc to dc, so that the roboteq doesent see the spikes from the contactor, or the chairs brakes, directly and only has to drive the SSRs. Low load/current.

[snaker]

Thanks JMGarage, it's now clearer to me.

The latching relay already embeds 2 diodes inside (one for setting, one for resetting). So I can bypass the diode for the contactor safety pin (DOUT).

For the diode in series with the precharge resistor. What type of diode should I use? What's its amps rating, voltage?

[Burgerman]

Those will probably be for the inputs to take the spikes off the controlling device. Not ACROSS the relay. 2 different things. I used a couple of these. https://www.ebay.co.uk/itm/203783114685

[snaker]

I found the diode as in your ebay link. It's sold for less than 1 pound in VN But I am afraid that the stud will cause difficulty when soldering it to a cable. I found this diode 20A10 (20A, 1000V). Is it Ok?

[Burgerman]

Should be, but 40A or 2 in parallel may be safer.

[LROBBINS]

Some further information:

The Roboteq Dout pins can handle a maximum of 1 Amp (and no more than 4.5 Amps for all Dout combined). I don't have its data sheet handy, but my recollection is that the latching relay John used draws much more than 1 Amp - hence using solid state relays between the Dout and a latching relay as he does is ESSENTIAL. Overloading a Dout will destroy at least that gate, and may even destroy the whole Dout chip. If just one gate goes bad, you can switch to using a different one, but if the chip is destroyed so is the Roboteq (unless you pay dearly to have it repaired).

Non-latching relays draw much less current; the one I'm using draws 0.17 Amps so can be powered directly from a Dout, but, of course, it is on the whole time the motors are powered. To improve its service life, the script has a CONTACTOR subroutine that opens and closes Dout for the contactor only when negligible current is flowing: a few msec before sending power to motors, or when motor output has fallen to near 0. This means that in normal operation the relay coil is not active all the time, but only when the motors are running and by doing the switching when little or no current is flowing means that there's no contact arcing to damage the relay itself. (This way of doing things was vetted for me by TE application engineering.)

Having the relay Dout switch on and off, however, means it can't be directly used for Roboteq's fault detection, so I also added a CK_RUNAWAY subroutine that actually does two things. One is that it reads the MOSFET fail flag in the Roboteq firmware and uses that to open the contactor Dout. The other is that it will detect a runaway condition and open the contactor in that situation as well; a runaway state is defined as anytime that motor command and motor power are ZERO but there's even minimal current flowing for 500 msec. These two subroutines essentially mimic Roboteq's fault detection, but still allow the CONTACTOR subroutine to work the relay without contact arcing, and not have its coil powered even when not moving.

The stud of a stud mount diode doesn't get directly soldered. Solder a ring terminal in your wiring harness and then use a nut to secure that to the diode.

Paralleling two diodes may not get you double the current rating. No two diodes will be exactly matched and the forward voltage drop can be substantially different for the two in at least some conditions - especially as they are warmed by current flowing through them. The result is that
one will pass more current than the other and may fail at not much more that the rated current of just one diode. What can happen is that first one diode fails, then all the current goes through the other one, and it fails too. Its better to use a single large capacity diode for the contactor's (reverse voltage) bypass.

[Vitolds]

I use a similar contactor switching circuit as here https://www.zeva.com.au/index.php?product=110
I also made a controller board with a mini relay. when the 12V power is turned off, the relay always turns off the controller.

[snaker]

I found this diode RHRG75120 (Fairchild 75A, 1200V). 75A seems overkill

[snaker]

I gathered all needed components around the emergency relay. Before moving to the next step, please help me to check if the diagram below is correct? I have a question about the precharge resistor. It's only 500Ohm/5W. Is it too small? Because as we expect, the current going back to the battery can be a dozen Amps. Can that current make the resistor melt?

[Burgerman]

The diode and the resistor go in parallel. You have them in series. The DIODE ALONE puts the regen current back into the battery should the fuse/relay trip. The resistor goes the opposite way! It allows the capacitors to charge inside the roboteq to prevent the contacts burning when he relay makes contact. Nothing to do with returning power to the battery.

And at 48V and roboteq at 0v with 500Ohm you will get this:


Current (I) = 0.096 Amps

Power (P) = 4.608 watts

Peak. Dropping further and fast over time. So no problem.

[snaker]

Thanks BM. I modified the diagram as bellow. Is it now correct?

diagram.png (10.98 KiB) Viewed 1983 times

I also ordered 3 Fotek relays (for the emergency relay, brakes and lights).

[Burgerman]

Yes. Relays OK but still needa TVS or diode on the roboteq input/output for safety.

[snaker]

Are TVS diodes fitted at the positions (marked by red rectangles) as in the attached image? Does their voltage match the max battery pack voltage? E.g
14s pack, max voltage = 14x3.600V = 50.4V => 51V TVS diode?
16s pack, max voltage = 16x3.600V = 57.6V => 58V TVS diode?

[Burgerman]

No. They prevent the brake, relay, lights control signal pin that feed the ssr's from seeing any spikes. So across the relay input connections. Presumably you will use the 5 or 12V supply for that? I dont remember now. And so a 15V TVS.

[Burgerman]

TVS is basically like 2 zenner diodes in one package. They allow current to flow in either direction but clamp it to a max V of whatever value you choose. So it protects the fragile control pins that you use to sink a small current in order for the brake/light/contactor to operate. You may get away with no diode or TVS by using SSR but better to be safe than sorry.

[snaker]

I drew a diagram of how I will wire the safety contactor pin (DOUT3). Please help me to check if the 15V TVS diode is fitted at the correct position?

[Burgerman]

I think so... I cannot follow diagrams, I just find it easier to do it in actual material parts!

I think I used the 5v on the roboteq to drive the ssrs, and the 12v fed the other side of the SSR. And then to brake, contactor etc. This stuff all gave me a headache at the time! At leasy your diode is the right way around. I aturally fitted mine pos end to the battery. Then it came to me while in the pub that when power was cut the robeteq fed the battery!

You need to think of every eventuality.

Have you made up any current sensors for the motors? It can be done without. But its not as precise. I use them.

https://www.wheelchairdriver.com/BM3-co ... -parts.jpg
https://www.wheelchairdriver.com/BM3-co ... r-rear.jpg
https://www.wheelchairdriver.com/BM3-co ... rimped.jpg
https://www.wheelchairdriver.com/BM3-co ... shrink.jpg

Or your RC setup?

Or your joystick? It needs RC on off latched switch, speed pot, inc end resistors, status LED, RC ON LED, roboteq on/off, contactor on/off. Etc.
https://www.wheelchairdriver.com/BM3-co ... -small.jpg

[JMGarage]

I think You don't need that diode for SSR. Diode is more for "old" style coil and loads.
Load side You can add that diode. Input side don't need SSR.
https://electricala2z.com/wp-content/up ... C_Load.png

[Burgerman]

Maybe. But it cant hurt. Its an electrically noisy place. It garantees that no spikes exceed the roboteqs in/out voltage limits.

[snaker]

I modified the diagram and put the TVS diode at the contactor's reseting pins. I also use 5V pin to drive the SSR. Is the diagram correct?

I am still at the initial steps. There is a long way in front.

I must test the project without current sensors first. I will not use RC. I will use an used VR2 joystick first for testing. I might upgrade APEM joystick later if everything is tested Ok and works. But those final steps are in a far future

[Burgerman]

The TVS goes across the ROBOTEQ 0V and the Dout pins to protect the in/output. It can actually go across the SSR control input side, as thats the same thing. It prevents spikes from reaching the SubD connector internal control chip. NOT across the contactor.
The LARGE one way diode goes across the contactor along with the precharge resistor.

And the contactor will also need a switch on and off (momentary switches at 12V) on the joystick.

[Burgerman]

I will not use RC.

Trust me it makes setting everything up, and figuring everything out 100x easier. And safe. And its also useful afterwards. You certainly will never regret it.

[snaker]

I move to the next step, the on/off switch. The wiring is something like in the image below. Is it correct?

BM, do you remember the value of the resistor in the diagram?

[JMGarage]

Wasn't that precharge resistor and should be over switch as in manual? 100R 5W. As earlier pic.

[snaker]

No, this is SW1 for MCU (not SW2).

[JMGarage]

Ah, You mean that switch. So pull down resistor.