EE question: convert RS232 voltages to RC servo

[Ether]

Looking for EE's to comment please.

I want to convert a -12V to +12V PWM signal to RC servo voltage levels (0 to +5V), with sufficient current to reliably drive a Victor.

I cannot locate any specs for the Victor's input impedance and required minimum input voltage.

I'm thinking a simple circuit like this will do the job. The questions I have are these:

- Is 1200 ohm a small enough output impedance to reliably drive the Victor's input?

- Is the 2.7MegOhm small enough to allow the 12V input to drive the 2N3904's base?

- Will the transistor operate as desired (i.e. shut off) with a -12V signal at the base, without damaging the transistor?

- Will these resistance and voltage values keep the 2N3904 within a safe operating range ?

[EricVanWyk]

Victors are opto-isolated, so really you just need to get your current in to the right range. You should be able to do this with a resistor and a(n optional?) diode (to be nice to the LED in the opto at -12V.

I'm not sure what value resistor would be best.

Your base resistor seems a bit high by my math, I'm not convinced it will turn on enough to register as a low... but my BJT math is extra rusty. I'd replace the BJT with a FET and lower the gate resistance to 100k, like the digital-age/class-D EE whippersnapper I am.

[Ether]

Am I interpreting your suggestions correctly?


And, does anyone have specs on:

• what minimum input voltages the Vic requires in order to operate correctly, and
  
  
• what input voltages the Vic can tolerate without damage, and
  
  
• what is the input impedance of the Vic?

[Hugh Meyer]

Ether,

As Eric said the input is an opto isolator. If you look at the schematic for the Jaguar, it is very similar, but not identical. A few years ago I made a device that would drive the victors and found it required about twice as much current as the Jaguars. I wanted the same device to drive either.

On your drawing you didn't say if the other side of the input was going to + vcc or gnd. If it is going to gnd, then 1200 will not supply enough current.

I will try to dig up the stuff I used. Seems like I remember it needed to be in the 400 ohm range.

-Hugh

[Hugh Meyer]

Ehter,

Check out this thread, if you had not already.

http://www.chiefdelphi.com/forums/sh...ghlight=victor

Hugh

[Ether]

Originally Posted by Hugh Meyer

http://www.chiefdelphi.com/forums/sh...ghlight=victor

Thanks for the link, Hugh.

In that thread, Don Rotolo posted:

Originally Posted by DonRotolo

you can use a 5.1 volt Zener diode (and a low-value series resistor) to clamp the input voltage to 5 volts or so.

Like this? That seems to me to be the safest way to proceed, since my concern is to make sure I do not damage the Victor:

• As long as the zener doesn't burn up, it protects the Vic by limiting the positive voltage to +5 volts applied at the Vic's input
  
  
• The zener also limits the negative voltage applied to the Vic input to approx -0.7 volts instead of -12V (as Eric suggested)
  
  
• The value of resistor R1 becomes less critical. It just needs to be large enough to prevent the zener from burning out, and small enough to assure that the voltage drop across it doesn't drop the Vic input voltage below 5 volts. Since Hugh indicated that 1200 ohms series resistance worked well with an 8-volt supply, I'm guesstimating maybe 1800 ohms for R1 would be a good choice for my 12 volt signal?

[EricVanWyk]

I suggest a fully series circuit: +/-12V -> diode -> resistor -> PWM input -> return.

Lets take a look at the jaguar because we have a schematic available for it. Its PWM input stage is a H11L1M in series with 150 Ohms.

The H11L1M has a threshold current of 1.6mA, and can handle up to 60mA continuous. It specs forward voltages at 10mA in the 1.2 to 1.5V range. It can only handle up to 6V in reverse.

Lets target 3mA. The forward voltage is in the 1 to 1.3V range (Figure 6 plus slop), and lets say that our series diode's forward voltage is .3V. Total silicon drop is somewhere between 1.3V and 1.6V, lets call it 1.5V. Our resistance should be (12V - 1.5V) / 3 mA = 3.5kOhm

Zener-ing it can be made to work, but that is more appropriate for a voltage-mode input. The Jaguar and Victor both have current mode inputs.

[Ether]

Originally Posted by EricVanWyk

I suggest a fully series circuit: +/-12V -> diode -> resistor -> PWM input -> return. ... The H11L1M ... can only handle up to 6V in reverse.

Doesn't there need to be a charge-drain resistor R2 to prevent reverse voltage across the H11L1M diode when the signal is -12 volts?

[EricVanWyk]

Originally Posted by Ether

If I put 12 volts reverse voltage across 2 diodes in series, won't I have 6V reverse voltage across each one? That's awful close to the max allowed for the H11L1M.

Shouldn't there be a resistor R2 across the H11L1M inputs to prevent this?

This is where things get painfully handwavy, because we are outside of what the datasheet specs. Also, you can't really assume that 12V/2 = 6V. You can think of diodes in series as a resistor divider where the resistors are extremely non-linear and suffer huge part to part and thermal variation. The math gets ugly and full of conflating variables. Make conservative assumptions and eliminate as much noise as possible, expect an answer with >=1 significant figure

Take the 1N914, selected by sorting digikey by price and choosing the first one with a decent datasheet. Figures 1 and 2 show reverse bias current curves, and are in the 20nA range for 12V. I'm confident the LED can handle this 20nA of reverse bias because at 3V in reverse it is leaking up to 10uA. The 6V number on the H11L1M sheet means that 6V @ zero ohms is being applied and survived forever in a worst case manufacturing / temperature scenario.

Another option is to put the diode antiparallel across the LED. The disadvantage is that it conducts in both directions. The advantage is that it is easy to prove from the datasheets that everything will be ok.

[Ether]

Originally Posted by EricVanWyk

Another option is to put the diode antiparallel across the LED. ... The advantage is that it is easy to prove from the datasheets that everything will be ok.

That's kind of what I had in mind when I drew Figure C for post #3.

Hugh mentioned in an earlier post that the Victor requires about twice as much input current as the Jag. So make the Vic 6ma. Divide that into 12 volts to get approx 2000 ohms for the resistor in Figure C.

Other than mail order and Radio Shack, are there any other places any more that might carry small electronic parts like resistors and diodes? For example, would RC hobby stores carry stuff like that?

I used to have drawers full of parts, collected from old radios and TVs and hi fi equipment I disassembled. Somewhere along life's road and many moves they got lost or discarded.

[MrForbes]

Originally Posted by Ether

Other than mail order and Radio Shack, are there any other places any more that might carry small electronic parts like resistors and diodes? For example, would RC hobby stores carry stuff like that?

not likely. Digikey is your friend.

I used to have drawers full of parts, collected from old radios and TVs and hi fi equipment I disassembled. Somewhere along life's road and many moves they got lost or discarded.

I know where mine went....the kid in college taking Computer Engineering has them. And he actually builds stuff.

[Ether]

Would this work? No +5V supply required. R1 = 1200 ohms, R2 = 5600 ohms. I have these parts on hand.

My main question is, would the +/-12V on the base turn the transistor on and off without damaging it?

[Ether]

The PWM outputs from the digital sidecar: can anyone tell me what their output impedance is please?