Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: MPower on November 08, 2014, 10:11:21 PM

Title: 24V Signals
Post by: MPower on November 08, 2014, 10:11:21 PM
Hello,
I am trying to use some servo drives that require 24V instead of the LPT 5V stuff. Is there a simple way to convert the step/dir to 24V signals?
Thanks
Title: Re: 24V Signals
Post by: GBowes on November 09, 2014, 06:10:42 PM
It should surprise me if someone doesn't already make an interface board that will do this.

A SN7407N IC (Hex Buffer/driver with open collector) might serve as a basis although I am sure there are newer, faster chips that might be better. Changing the signal levels should be fairly easy but you would have to be cognizant of the implications on timing. The 7407 is a very old chip and has average propagation delays of 14 nanoseconds.

Unfortunately, unless you can find someone already producing an interface board, you would have to design the complete circuit around your selected chip, although that should not be too complicated.

Graham
Title: Re: 24V Signals
Post by: MPower on November 09, 2014, 06:27:16 PM
Yes, there are a couple of similar chips that act like that, however I couldn't find any exact implementation. I assume just putting the chip between the 5V from the LPT breakout board and the servo drive that expects 24V instead of 5V, would not be the most reliable way to do it.

According to the servo drive datasheet, the pulse width should be >= 2.5us and pulse pause >= 2.5us.
Title: Re: 24V Signals
Post by: GBowes on November 09, 2014, 07:00:35 PM
I did a quick Google and came up with this. I only took a superficial look but it might do what you need.
http://www.talent-hk.com/data/TEK6%20DatasheetR2.0.pdf

I assume just putting the chip between the 5V from the LPT breakout board and the servo drive that expects 24V instead of 5V, would not be the most reliable way to do it.

According to the servo drive datasheet, the pulse width should be >= 2.5us and pulse pause >= 2.5us.

This is essentially how to do it. you would need a 24VDC supply, some pullup resistors and probably a capacitor or two on the power supply circuits to help avoid noise. It might also be a good idea to opto isolate the inputs from the parallel port. In theory it would not be complicated. In practice it will be all about timing errors and pulse shape. If it were my machine, I would probably rough something up and see how it works but tweaking it to work reliably (if possible) would require hands on analysis with at least a DSO to study the timing.

The 2.5uS pulses would be no problem. You can set them in Mach3 and the IC would reproduce them within its timing constraints.

Graham
Title: Re: 24V Signals
Post by: MPower on November 09, 2014, 07:17:37 PM
Well :)
The motors I am trying to use are Berger Lahr VRDM 3910/50 LWB (VRDM 3910 LW 4 Nm 230 V 2,00 A). They are driven by Berger Lahr WD3-007.0501 (WD3-007.0501 230VAC 24V-signal-input). The drivers were signaled by a Siemens PLC.
The driver in the link will not work.

There is a 5V signal version of the Berger Lahr (WD3-007.1501 230VAC 5V-signal-input) but the costs for the buying 3 units for 3 motors is ridiculous. I would rather buy modern servos at that price, the WD3 was released somewhere around 2003. The only chance would be to bump the LPT 5V signals to 24V, in theory even optocouplers would work. I was hoping somebody already went through this (being a common problem if you wish to use 24V industrial stuff via Mach3 and LPT).

Also, the specs of the driver:
Uhigh, min 20 V
Uhigh, max V 30 V
Ulow, max 3 V
Ulow, min -30V
Imax = 30 mA
Current Control
Ihigh, min 7 mA
Ihigh, max 15 mA
Ilow, max 0.2 mA
Ilow, min -25 mA
Umax = 5.25 V
Title: Re: 24V Signals
Post by: TPS on November 10, 2014, 03:01:35 AM
Hi,

have a look for CS-Lab controllers, they are on 24v industrial standard.

Thomas
Title: Re: 24V Signals
Post by: MPower on November 10, 2014, 06:10:15 AM
Hi TPS,
I assume you are referring to this product:
http://www.cs-lab.eu/en/produkt-21,4-Universal_converter_of_digital_signals_with_optoisolation__5V_input_voltage.html ?
Title: Re: 24V Signals
Post by: TPS on November 10, 2014, 07:21:22 AM
ok,
sorry i was wrong, your berger Lahr controler has also 24v for step/dir signals.

than you have to look for optocouples like this:

https://www.phoenixcontact.com/online/portal/us/pxc/offcontext/search_result_page/!ut/p/b1/vZPJkqJAEIafpR9Aq4pVjsXiglKyFApcCBoLFWRRaUGevu2ZiL51e5gJMy-ZEV_GH1WZP4hAiGRBViYiEgUQgKhKbsd90h7rKjl99ZEUU5Uahs1JaO2IKlzoomasBHcOefQAwh8BNHf4f5vH6Nl8-ADkWJkRCXKChGaiokPM-ZrtrXzdwxKgIIBC7OX3ZjEUg5sP7tqiFiTU6BGEiOT29lEPZKcSoicdNDhEdJ8jEN69B-sp1N5tXF_FGtlBa_ZMUHyxoPPqF0rCqwX__5eaIDq-l-MuLcdw3PTpuM6ytK5a1rfjK0su6SG-sOvHqY2bZM_AJgivsl5bnYED27lNE6X1Rc0v6OzdMkxbKVbDUsySRT3dcZZ7XFNeZ0O6t1X-9iEM_ByuQ7PSl6XasHAamJ3I-nSa-VJ5Z7TfK6TA9aiUgttFX0codlh3TtoRQ1HrysfzgaRdhrVBC6gPtiD6xRB__fYr8GWoPwD8ITAEHqti34s3tu8BMq9L9uzm5e8FTS7WKs1NagnXSYpsL-Xd_N41y4LfqF1dmc2uED-yxu2TyjyEp2iUV9vGueC95p1zHYEyOk0n6Dv5xsgwxG9vn9Put9k!/dl4/d5/L2dBISEvZ0FBIS9nQSEh/pw/Z7_9GN602461G59D0A2UCPSLUDSA7/ren/p=pxc.param.sort.field/p=pxc.param.pager.requested.row.count=10/p=pxc.param.pager.selected.page=2/p=pxc.param.sort.direction=ASC/-/?uri=pxc-oc-itemdetail:pid=2964270&library=usen&tab=1#Connectiondata


Thomas
Title: Re: 24V Signals
Post by: MPower on November 10, 2014, 07:46:32 AM
Looks like it could handle the task. Did you use this or similar SSR for converting the signal voltage? I am not sure if I just plug the LPT on one side and the Berger driver on the other side and that's it, or I have to add some extra filters for the signal quality improvement.
Title: Re: 24V Signals
Post by: TPS on November 10, 2014, 08:00:16 AM
hi,
i think on the 5v side its only to connect to lpt port,
but on the 24v side you need to add a power supply,
because the berger lahr is only a input.

Thomas
Title: Re: 24V Signals
Post by: MPower on November 10, 2014, 08:02:20 AM
Yes, the 24V side will be powered by an auxiliary 24V power supply. It's just that it looks quite simple, I was expecting a far more complicate solution :)
Title: Re: 24V Signals
Post by: GBowes on November 10, 2014, 06:46:47 PM
Hi,
As I mentioned before, it is quite simply in theory.
That link to the SS Relays didn't work for me but looking at some of the products on that web site, they could work. I see they have special configuration relays that can run to 100 kHz which is probably more than you need.
It appears that the standard relays are rather slow and this may cause problems but I also think this approach may be just as simple or complicated (depending on point of view) as using a buffer chip. The only thing you save with the relays is the pullup resistors/

The more I think about this the more I think my original suggestion of the SN7407N type chip would work as well or better and be a lot cheaper: ($1.00) for the chip that can handle step and dir for 3 axis, 2 pullup resistors per axis, and a couple of inexpensive capacitors on the 24V power supply.
I'd also recommend isolating the PP from the 24V with more than just the 7407 or the relays and would use HCPL2531 optocouplers These are very fast and are only $2 each with one per axis .

I did not think it through carefully but I also believe the CS Labs Universal Converter (the 63 Euro one - differential TTL to OC) could handle everything if configured as an open collector output. Not inexpensive but almost a complete solution rather than a bunch of parts.

Graham
Title: Re: 24V Signals
Post by: MPower on November 10, 2014, 07:27:14 PM
GBowes: here is a shorter version of the URL that TPS posted
http://bit.ly/1sx9sXo

(the long one seems to have been altered by the forum)
Title: Re: 24V Signals
Post by: GBowes on November 10, 2014, 07:54:40 PM
Thanks. That relay would be great solution except for one slight problem. It has the isolation and buffers the signal and converts to 24V. All you would need is the 24V supply and a couple of capacitors, You would need two relays for each axis (step + dir).

I had guessed that it might be a little expensive, but I was wrong. >:(
Mouser wants $92 each so $560 for the 6 needed for three axis.

That makes the CS Labs unit look cheap at about $80 plus shipping.

Graham