3v logic -> high side FET switch

21 posts / 0 new
Author
Message

I'm trying to interface the logic signals from a RF receiver, RXD-418-KH2 (data sheet below) that runs at 3v to a variable frequency drive, model ATV11 (data sheet below) digital inputs that expect 15v nom. It's inputs have a 5kohm resistance (I assume to ground) and it provides the 15v. see page 28 of data sheet. The VFD needs >11v for logic high, and <5 for low.

I don't know the best way to do this, I try to search for suggestions, but I'm wasting time spinning my wheels. I would like a monolithic solution that would take the +15v rail, a 3v logic input and produce a non-inverted output.

INPUT | OUTPUT
-----------------
0v | OPEN or 0v
3V | 15v

A hint in the right direction would be great.

Also, since the VFD is driving equipment it is important that if the driver fails it fails open so the machine doesn't continue running on it's own.

Thanks,
Ben

RXD-418-KH2 data sheet:
https://www.linxtechnologies.com/resources/data-guides/rxd-xxx-kh2.pdf

ATV11 data sheet / user manual:
http://static.schneider-electric.us/docs/Motor%20Control/AC%20Drives/Altivar%2011/VVDED302026US.pdf[/img]

Oh, I was going to mention solutions that I bounced around.

I considered using an inverter and BJT to drive small relays, (3 channels needed). That cost added up and seemed overly complicated.

Also just using an inverter and BJT with a resistor for a load which would be the top half of voltage divider, and the 5Kohm impedance in the VFD would be the low side, so that would require ~1250 ohm "pull up" resistor to get 12v to ensure logic high. Again, no experience, so not sure if it's a good way. Should work though.

Have you seen the circuit collection on ruggedcircuits?

Imagecraft compiler user

@ Mr. Gardner,
Looking now, thank you. I like it.

Strange, I never find what I am searching for when I cannot see the forest from the trees. And I tend to search in the wrong forest... :roll:

2 transistor type PDTA144ET where E is to 15V input and C to your 15V output.
and a PDTC144ET where B of PDTA144ET to C of PDTC144ET
E of PDTC144ET to GND and the B of PDTC144ET to the output for your RF circuit.
this transistor has inside bias resistor. not need extern resistor.
you can find one SOT23-6 with this both transistor inside if you have place problem.
Thierry

Thierry Pottier

What happens if the receiver fails? Don't expect a high degree of reliability - especially if safety is concerned.

Quote:
Kartman
What happens if the receiver fails? Don't expect a high degree of reliability - especially if safety is concerned.

I worked up a "chart" if you will, of failure modes and the problems that I can foresee are:

Biggest one: poor RF link. But these should all cause a "no motion" failure. Which is little more than a nuisance.

My biggest concern was a "false on" where for example a relay would stick closed for some unforeseeable reason. I thought sticking with solid state over mechanical switching would be better in regards to that. Could be wrong though...

As for the receiver failing, it is surely possible, but since it was not built by me and is an OEM solution I kind of assumed it would not be the weakest link. Again perhaps a bad assumption, but I didn't find any suggestion in the data sheet where the failure would result in a false true output.

My last concern was that with my own devised switching circuit, I would inadvertently (again lack knowledge about physics of transistors) create a situation where they might "latch up" due to transients at power up or something. Going with circuit examples like on ruggedcircuits.com, (thanks Bob) keeps me using tried and true methods. I guess I've past the point in life where reinventing the wheel sounds like a fun thing to do.

@ TPE:
I put your suggestion in Eagle, using MUN2111 and MUN2211 resp. (didn't have your part numbers in any libraries) and it looks similar to this:
http://ruggedcircuits.com/html/circuit_-19.html
only with extra resistors and no fly-back diode. One at ground to the NPN base and the other at +15v to PNP base. I don't know what purpose they serve other than allow extra current to flow. Having the whole solution in two packages >is< nice.

Now having an idea what to look for, I search digikey and get to:
http://www.digikey.com/product-search/en?pv74=98&FV=fff40015%2Cfff803f3%2C1c0002&k=dual+npn+pnp&mnonly=0&newproducts=0&ColumnSort=0&page=1&stock=1&quantity=0&ptm=0&fid=0&pageSize=25
now I just need to make sure that the R1 and R2 so that a logic 3v is suitable to drive the input. :shock:

Thanks for your help guys.

TC4427

Science is not consensus. Science is numbers.

I don't want to harp on the subject, but if your device could cause injury or death, then your assumptions are invalid. If there are no safety concerns then your setup ahould be adequate.

@ Kartman,
Harp away, no harm in that. Safety should be at the forefront. That is why I am looking for the "best" way to switch the signals. Got to thinking of the projects and gizmos I've worked on, electrical and other, nearly everything has safety concerns.

Obviously, there is an E-stop button on the enclosure at the Rx end that cuts main power supply. It's also just powered by a 120v cord so unplugging is always an option, but not obvious as a red button if someone in a panic.

The equipment that I'm controlling with remote does have the ability to cause injury and possibly death, but not a whole lot more than a garage door opener. The speed of the motion is limited to about that of an automatic garage door. Only we're rolling large tanks (3k to 10k gallons) empty of course. That on it's own sounds dangerous, but it is a necessity in fabricating. Going to remote has many advantages over a cord that make it safer, but it does have drawbacks as well.

the choice of resistor depend of output current.. but you need low current, Hfe for this transitor er better than 100 and if it is only on/off not high speed switching, you can hold low base current.
http://www.digikey.com/product-d... is nice!
I use one dual Npn/Pnp like this but to hold power for one project.

Thierry

Thierry Pottier

Did I miss something? You said you wanted a monolithic solution. Have you looked at the data sheet for the TC4427?

Science is not consensus. Science is numbers.

or you could try this ULN2003A
im using it to drive 12v relays with 3.3 logic

Apologies hobbss, I got distracted with reconsidering failure modes and safety. But, I did look and the data sheet and I think the TC4427 should do it. I like the line:

Quote:
These devices are highly latch-up resistant under any
conditions within their power and voltage ratings.

Also, if I read correctly anything over 2.4 volt at input should turn it on, so it will work with the 3v logic level, right?

Thank you,

Ben

These things were designed to drive high gate capacitance fets (think H-Bridges) - hence the large current drive capability.

Check the lowest Voh on your driver (see data sheet) to make sure it will meet the minimum Vih of the TC4427.

Science is not consensus. Science is numbers.

If this unit is going to be used in the workplace, you have oh&s to consider. In soing a failure matrix, all your points are likely to fail and the outcome is "not good". You need to mitigate potential failure. The usual rule is no one failure will cause an dangerous outcome but the failure can be detected. The transmitter would need to ensure the buttons are working and some form of interlock so it could not accidently transmit. The receiver would need to receive changing data and toggle output bits to guard against stuck bits. That would feed a watchdog circuit with conservatively rated components. That output would need to be checked and worked into an emergency stop circuit.

I can tell you from experience that things do fail in unexpected ways. I had a batch of boards with counterfeit relay drivers. These would slowly degrade over time and hold the relays on. My circuit detected this and failed safely. The fault was recorded in eeprom. After a few reports of failure in the field, I was able to identify the fault and determine the batch that was affected.
The circuit was designed to tolerate relays being stuck on - all on and all off were safe states. i had a failsafe relay on a discrete watchdog circuit as a catchall.

I have a fault of over complicating things and adding to many "features" that usually just end up causing more pain than gain. So I tried to apply the KISS principle to this. Thinking the simpler solution = fewer ways to fail. You are convincing me otherwise. I hoped to keep microcontrollers out of both ends, just didn't think they were needed and wanted to limit number of things to go wrong. I remember in computer science classes, every time I added a "feature" to a little program it seamed I added two bugs. But I like programming microcontrollers, that's fun.

I like the concept of always expecting changing bits. I think I could do it with logic gates, but at that point a uC sounds like a quicker solution with no down side. I'm going back to the drawing board, bummer, I had the RX board all routed...

As I consider the way the rolls come OEM, I don't know how they could consider them safe for the environment intended. The biggest problem is the 30 foot multi-conductor cable that carried signals (120VAC) from the operator station (hand held box about 5" X 5" X 20" steel enclosure) to the enclosure that has the VFD in it.
1) We fabricate tanks out of sheet metal, mostly 12 and 10 ga (~1/8") thick. Sheared sheet has sharp edges, cords are continuously hung over edges of steel. When tank is rolling and operator doesn't see cord getting pulled tight, insulation gets cut, sometimes both the jacket and inner insulation. I don't need to explain more.
2) I'm in the process of switching out steel enclosures for with plastic. The metal enclosure needs to be grounded for protection. Problem is if the enclosure touches the tank while welding (very normal) and the welding ground clamp disconnects (common as well) the full welding current tries to travel back through the 14awg conductor turning it into a 30' long fuse and making smoke out of the insulation around its neighbours in the cable.

In my 4 years there I've seen both these happen (and replaced the cords) three or four times. Maybe more, I forget.

BTW, Kartman, sounds like you've been at this a while?

From what you're saying, there are a few problems that have been worked around for too long. You really want a metal enclosure for the vfd for rf shielding. You should have special braided cable to the motor. To mitigate the earthing issue, provide a more substantial earth path for the errant welding currents.

As for the remote control, the wired method is the simplest. Maybe run the cable in flexible conduit so it doesn't get cut as easily. Personally, i'd suggest purchasing a remote control if you want to go down that path. There are ones available for overhead cranes that are safety certified. Not cheap though. You'll probably have to make a business case to highlight the potential for higher productivity and increased safety.

This works against your goal of doing a remote control project. I'd say the potential risks outweigh the benefit.

In many industrial accidents, the root cause comes down to people circumventing bad processes or equipment. Once you lift the lid and look closely, many of the solutions are surprisingly simple. Would having a remote control solve the root cause or make a bad process easier?

As of last year i was involved in automation and instrumentation for oil and gas industry. Unfortunately, pointy haired managers drove me away.

bnemec wrote:
Got to thinking of the projects and gizmos I've worked on, electrical and other, nearly everything has safety concerns.
Where the gizmo resides (industrial, railway, etc.) can show which safety standard to use.
Leading safety critical standards and safety integrity levels. (EE Times)
Atmel created some AVR code for an appliance safety standard.
Atmel AVR1610: Guide to IEC 60730 Class B Compliance with XMEGA

"Dare to be naïve." - Buckminster Fuller

Quote:
From what you're saying, there are a few problems that have been worked around for too long. You really want a metal enclosure for the vfd for rf shielding. You should have special braided cable to the motor.

VFD in metal enclosure, it's the box with the buttons at the far end of the cable that should not be metal. Cord from VFD to motor is about 18" long, motor is right behind VFD enclosure. That part of the rolls I'm not touching, that's all original from factory.

Quote:
To mitigate the earthing issue, provide a more substantial earth path for the errant welding currents.

The tank is moving a lot, the ground wire gets repositioned frequently. Not sure what to do here. If there's any welder that can truly say he/she has never forgotten to hook up the ground, I would have to doubt they've been welding very long. Thinking about it, removing extra ground paths >is< the process improvement. The guy that worked on these before me had a solution, he just ripped the grounding prong out of the power cord, then they couldn't ground out. Problem then the whole roll wasn't grounded. I am trying to make improvements.

Quote:
As for the remote control, the wired method is the simplest. Maybe run the cable in flexible conduit so it doesn't get cut as easily.

I'm poorly describing the typical processes used in fabrication of these tanks. The wired method is certainly the easiest to implement, but it's a pain to use with very high maintenance requirements. Of the dozen tank rolls in the shop I estimate a cord gets the jacket cut, stretched or melted every week. As for they labour force, they get what they pay for. I'm not in management.
Putting something like Anaconda on the control cord would render it useless. Perhaps a woven fibreglass sheathing like McMaster sells would work, I use it on TIG torch leads, but it's not certified for anything.
The length is becoming an issue too. We're starting to build tanks that are 30 feet long. The control box needs to go inside the tank with the worker, through the man way. These would need 60' long cords to be able to reach both ends of the tank and allow it to roll all the way around.

Quote:
Once you lift the lid and look closely, many of the solutions are surprisingly simple. Would having a remote control solve the root cause or make a bad process easier?

I'm not an Industrial Engineer, I don't know much about process improvement. If there are better ways trying to change the set-up is beyond my pay grade, I'm just trying to make things better, more bullet proof, for the guys with what I've been given to work with.

Thanks for the comments and suggestions, working through the problem is helping me rethink the solution. Again.

@gchapman, I'm looking up the link...

Ben

bnemec wrote:
The wired method is certainly the easiest to implement, but it's a pain to use with very high maintenance requirements. Of the dozen tank rolls in the shop I estimate a cord gets the jacket cut, stretched or melted every week.
Likewise as a part-time operator and full-time engineer.
The remote's cord would sometimes get run over and almost always stepped on.
A different (non-remote) relatively high pin count connection would fray its shield and the pins/sockets would look well worn.
The work around was a thicker plastic jacket so a heavier cable and some grunt to mate and demate.
I never operated in polar vortex conditions but I imagine difficulties due to the required insulated gloves and insulated coveralls.
RFDANT (RF Digital, antenna 2-way radio) might be rugged enough, some security and reliability with built-in CRC and their protocol, 9600bps, long outdoor range.
If their antenna radio could get sheared off when dropped then their modules may be an alternative; module has a wire antenna app note that may fit inside a large handheld ABS (or stronger) plastic case.
The module cannot be potted but the data sheet states how to tent the module.
OpenSource RF is RF Digital's Arduino products.

"Dare to be naïve." - Buckminster Fuller