Shopbot PRS 48x96 Acorn / Clearpath Conversion

[Richards]

The IPC-5 is rated at 500 Watts. The SDSK-3411 is rated at 220 Watts peak and 203 Watts continuous. You'll be using four motors. In your case, with a 1:5 gearbox, and at speeds that Gary suggested you won't be working those motors very hard. If I were designing a new enclosure, I would consider leaving room for two power supplies and a Power-4 hub, then you could easily add whatever you might need. I look at a DIY project as a 1st or 2nd generation prototype where the builder has a pretty good idea of his needs but where the unexpected still happens.

[tkbot47]

Gary,
That is correct; however the hub can be connected to two Ipc-5 power supplies and four motors (more if multiple hubs are used). I've made custom cables and I've used stock cables. Crimpers to make good cables are pricey if they are efficient. Of course on a router custom length cables might be needed.

Mike,
I don't believe you can hook 2 IPC-5's to one hub, per Teknics. You can daisy chain motors or 2 hubs to one supply, but not the other way around, as I understand it. That would connect the DC output of two supplies in parallel.

[Richards]

Mike,
I don't believe you can hook 2 IPC-5's to one hub, per Teknics. You can daisy chain motors or 2 hubs to one supply, but not the other way around, as I understand it. That would connect the DC output of two supplies in parallel.

I hurried and checked the ClearPath Manual, pages 36-42. You are correct in that EITHER input can be connected to a DC power buss. However, connecting the outputs of two identical power supplies results in the output voltage being the same as if one power supply were used, but the current capacity of the power bus is the sum of the currents of each power supply. I.e., one IPC5 power supply gives you 75VDC at up to 500 Watts and two IPC-5 power supplies gives you 75VDC at up to 1,000 Watts.

I have contacted Teknic support for clarification.

Meanwhile, I'm attaching an article from DigiKey that explains why using parallel power supplies can be a benefit.

[tkbot47]

Gary,
That is correct; however the hub can be connected to two Ipc-5 power supplies and four motors (more if multiple hubs are used). I've made custom cables and I've used stock cables. Crimpers to make good cables are pricey if they are efficient. Of course on a router custom length cables might be needed.

Mike,
I don't believe you can hook 2 IPC-5's to one hub, per Teknics. You can daisy chain motors or 2 hubs to one supply, but not the other way around, as I understand it. That would connect the DC output of two supplies in parallel.

I checked the IPC3-5 manual (pg 2) and it warns specifically as follows:

IPC Use Instructions
To connect your IPC power supply to a load:
• Disconnect IPC from AC mains power.
• Connect DC power cable from IPC to the load.
• Apply AC power to IPC.
To disconnect your IPC Power Supply from a load
• Turnoff(unplug)IPC.
• Disconnect DC power cable from the load.
Additional Use Notes
• Do not wire multiple IPCs together; they are not designed to operate in series or in parallel configurations.
• Always use recommended wire gauge (or larger) for all cables connected to an IPC power supply.
• Understand and follow all safety markings and warnings printed on the IPC and described within this document.

If you connect the outputs of 2 IPC's to the inputs of one hub4, since those inputs are connected to each other, you would essentially be wiring the 2 ipc's in parallel, which they warn against above. I would also be concerned about the hub4's ability to handle the output of 2 supplies, and since the ipc's employ automatic regeneration power management, I would be concerned about cross connecting outputs.

I recommend 1 hub per ps, the hubs are only $50

[Richards]

You may be correct about using only one IPC per Power-4 hub. I've contacted Teknic and am waiting for a response. I sent the support request late Friday, Eastern time, so they probably didn't have time to process the request yet.

If you used two IPC's to power four motors, you could hook up two motors per IPC without using a power-4 hub. Each IPC has two power out ports. Normally, if I suggested to use two power supplies in parallel, I would also recommend using a large diode in series with each power supply to keep voltage spikes from one power supply from traveling to the other power supply, but one of the features of the IPC power supplies is that they are designed to absorb voltage spikes when a motor is rapidly decelerated. Installing a diode in series with the output would prevent the IPC from functioning correctly.

Another option that costs much less than buying two IPC power supplies is to buy one linear power supply that uses a toroidal transformer. Antek sells the PS-8N70 that produces 70VDC at 800 Watts. It costs $140 plus shipping. The 1000 Watt PS10N70 cost $162 plus shipping.

I use a 500 Watt toroidal power supply on my test bench. It is connected to a Power-4 hub. I've used two IPC-5 power supplies with a Power-4 hub to verify operation to four ClearPath servos. The IPCs and one motor are currently sitting on a shelf until I build another controller.

[tkbot47]

I think that Teknic's declaration in their IPC-3/5 user manual not to parallel connect their ipc supplies would be an official statement from them. In any case, they should not be hooked together. I went with 2 IPC-5's and 2 hub-4's for belt and suspenders power, regeneration power management, as well as future options with the hubs. I have used Antek supplies before - they have lots of power, but are bulky, heavy, and not regulated, so I didn't want to use them for my Clearpath application. Very happy with the performance so far.

[Richards]

You are correct. Only ONE IPC power supply can be connected to a Power-4 hub. Tom, from Teknic, just sent me an email to let me know that only ONE IPC should be connected to an IPC.

That will make me rethink power supplies for ClearPath motors. Three 34xx motors for X, Y, and Z can pull more current than one IPC-5 power supply is rated to handle. If two 34xx motors are used as master/slave on a router, then they could require 800 W if run at their full potential. I could buy two IPC-5 power supplies or I could use just one 800W or 1000W toroidal linear power supply for much less than the cost of just one IPC-5. I've used linear power supplies for all my stepper/servo systems since the early 1990s on Kodak-S printers. None of those power supplies ever failed. They take up more room than one IPC-5, but are about the same size as two IPC-5 power supplies. They are very simple power supplies, just a toroidal transformer, a large bridge rectifier, a few large filter capacitors and a bleeder resistor to discharge the caps when power is shut off. They can easily absorb any voltage spikes caused by a rapidly decelerating servo.

[tkbot47]

I use 2 3432p, one 3421S and one 3421D between two IPC-5's on my Buddy redo, and have never had a problem. My concern with the unregulated Antec's are that the voltage droops quite quickly as current draw increases (check out the graphs on their website), and I wanted to maintain full voltage near max current draw, which the IPC's are supposed to do. I only wanted to spend once on power supplies, so went with the Teknic recommendation. It's probably a bit overkill, but I wanted to make sure the first time.

[paulwall]

After much debate, and after conversations with Gary, Tom, and a tech rep at Teknik I believe this is going to be the motor package I order.

Z: I'll be running 3421S-ELN as I intend to primarily run a lot of Z heavy 3D carving and may use a heavier spindle in the future than my 2.2hp HSD and want quick reaction times and movements here.

Y1, Y2 will be 3411S-ELN (Shopbot of course calls long axis of the 4' x 8' as 'Y'). The tech rep at Teknik with the parameters I shared says that (as Gary has confirmed) the 3411S should in theory be able to run ~1300-1400in/min with sufficient torque to provide 0.31G of acceleration (though Gary says 1000in/min is his max recommended speed before the table becomes unstable). The 3421S in this same application could push 0.64G of acceleration and still achieve around 900in/min, but this is where the concern lies. This speed is fine by me, but I share the same concern Gary has, that at this level of acceleration, the gantry only constrained by gravity may very easily skip off its rails. If this was a linear rail system, this would be a different conversation.

X - Still deciding between 3421S and 3411S as I have less concern for V-bearing rail jumps on X as they're more mechanically constrained than 'Y', but at the same time the mass of the carriage is far reduced. It's very likely I'll go 3411S here as well.

[Richards]

Your numbers look good. I just ran a spreadsheetfusing 1.25" diameter pinions and a 1:5 gearbox. I was more interested in continuous torque. The 3421S would have about 1,445 oz-in continuous torque with the 1:5 gearbox. That's a little more than the Shopbot advertised PRS Alpha torque. I had a PRT-Alpha upgraded to the PRS-Alpha motors on X and Y axes. The 3411S has about 750 oz-in continuous torque with a 1:5 gearbox. Comparing a servo's continuous torque to a stepper's holding torque is apples to oranges, but either motor looks good to me.