Railpower 1300 testing

[Doug G.]

Just wondering out loud here. Would a zener withstand, say, the typical user shorting the output and before any protection device opened the circuit? That may be a reason for the unregulated designs.

Doug

[peteski]

Just wondering out loud here. Would a zener withstand, say, the typical user shorting the output and before any protection device opened the circuit? That may be a reason for the unregulated designs.

Doug

The Zener diode is on the base (low-current) side of the power transistor and does not pass the heavy track current. It basically clamps the base voltage at 18V.  It is nothing like the series-connected Zener diodes Max uses to "waste" some voltage in his low-voltage coreless motor designs.  Those Zeners pass all the current going to the motor.

The only thing that I don't think will work in this circuit is that when the speed knob is adjusted to maximum speed, it connects the Zener diode directly to t he positive side of the bridge rectifier. Since the voltage there is quite a bit higher than the Zener voltage, the Zener diode will conduct rather heavy current.  It will be actually clamping the voltage coming out of the bridge rectifier!  There should be a resistor added to limt the current. Either between the wiper of the pot and the Zener's cathode, or between the top lead of the pot and the positive side of the rectifier.

[Doug G.]

OK, I guess I wasn't thinking about where it was in circuit. I am not a circuit design expert. Also, your comment on motor current draw. Yes, motors are not like typical static loads because of back EMF. Still guaranteed they follow the law, however.

Doug

[mmagliaro]

Peteski... good catch!

Yes, at maximum throttle, the speed pot path through to the zener is zero ohms, so the zener conducts maximum current directly from the rectifier.

Adding a 100 ohm resistor will fix things and have no noticeable effect on the speed control.  Out of 20k ohms,
100 ohms isn't going to matter to the top speed.  At full output, the current through the zener is now only about 15 mA.

Okay, so I need a zener and a resistor.  I still bet I'm under 50 cents.  Time to take this out of the circuit simulator and put these two parts into the actual 1300, and photograph where I put them on the board so people can see it.  I know I promised somebody I would do that back when this all started and I never got around to it.

[peteski]

Peteski... good catch!

Yes, at maximum throttle, the speed pot path through to the zener is zero ohms, so the zener conducts maximum current directly from the rectifier.

Actually I mentioned that design "flaw" when you originally posted this circuit. In this post. But I didn't really go into explanation.  And Point353 (in a post right above the one I linked to above) also hinted about the Zener diode current at full speed setting.

[Point353]

Okay, so I need a zener and a resistor.  I still bet I'm under 50 cents.

You should be able to get both components for about the cost of a cup of coffee at Wall Drug:

[peteski]

You should be able to get both components for about the cost of a cup of coffee at Wall Drug:

A quick peek at Digikey, shows that you can get a 20V 1W Zener for 6 cents in quantity of 2500. A 100 ohm 0.25W 5% resistor costs half a cent in quantity of 5000.  So it would be much less than a cup of coffee.    Single quantity will still be well under a dollar for both components.

[Point353]

A quick peek at Digikey, shows that you can get a 20V 1W Zener for 6 cents in quantity of 2500. A 100 ohm 0.25W 5% resistor costs half a cent in quantity of 5000.  So it would be much less than a cup of coffee.    Single quantity will still be well under a dollar for both components.

The implication was that the manufacturer of a power pack could buy (in production quantities) - as opposed to an individual (buying in unit quantities) -  both the zener diode and the resistor for a total cost of about a nickel, such that the incremental cost to include the two proposed voltage limiting components would be very small.
Nothing beats having one's words parsed by passel of peteskis.

[peteski]

Nothing beats having one's words parsed by passel of peteskis.

Meow!
Max's and your posts simply got me curious about the actual cost of those components, and since I looked them up, I simply posted my findings. I wasn't parsing anything you wrote. Relax, no need to get defensive.

[mmagliaro]

Well... I bought mine off eBay, so they came from China.  I got ten 20v 1W zeners for, I think, 2 bucks.
The resistor, well, I buy those 100 at a time off eBay and they are on there right now at 100 for $1.89.
So 19 + 20 = 39 cents.

Now, back to the throttle.
I put the mods in.  With a 20v zener and 100 ohm resistor, it works beautifully, exactly as planned.  Full load RMS track voltage is about 12.8, no load RMS track voltage is about 13.5, and the peaks are at 20v (under load or not) (instead of 25v in the stock MRC 1300.

However, wanting to do a bit better, I changed to a 220 ohm resistor and a 16V zener.  This drops the maximum peaks down to 17.5 with zero load and 16V under full load.  The RMS track voltage at full load is 10.8V, and with no load, about 11.5   
You could do it either way and you'd be fine.  If you want that last bit of extra "oomph", use the 20v zener. 

The 220 ohm resistor takes some load off the zener, at the expense of maybe 1/2 volt at the output at full load.
I did this because at full load, the 16v zener was getting a little hot.  I measured the current through it as 31mA, so .031 x 16 maximum = .496Watt, only half of the 1W rating for the device.  It's just that a half watt of heat is, well, hot.  It's fine as long as it doesn't touch anything.

Now, here are the scope plots and the circuit board mod.

[Point353]

I put the mods in.  With a 20v zener and 100 ohm resistor, it works beautifully, exactly as planned.

However, wanting to do a bit better, I changed to a 220 ohm resistor and a 16V zener. 

The 220 ohm resistor takes some load off the zener, at the expense of maybe 1/2 volt at the output at full load.
I did this because at full load, the 16v zener was getting a little hot.  I measured the current through it as 31mA, so .031 x 16 maximum = .496Watt, only half of the 1W rating for the device.  It's just that a half watt of heat is, well, hot.  It's fine as long as it doesn't touch anything.

If you're concerned about running the 1W zener at 0.5W dissipation, both Mouser and DigiKey stock the 1N5353 (16V) and 1N5357 (20V) 5W zener diodes, which cost about $0.50 each in single piece quantities. 
Also, you might not want to rest the body of the diode directly on the p. c. board.

[nkalanaga]

I'd go with the 20V Zener if you plan on running sound decoders on DC.  My Athearn Challenger will not run at much below 12V.

[peteski]

Max, what is the typical gain of that Darlington transistor, and what is the maximum current this throttle can supply?  I have a slightly different circuit in mind (which will dissipate even less power).

[MK]

Thanks for the circuit mod pictures.  I'm going to give that a try to my unit.

[mmagliaro]

Max, what is the typical gain of that Darlington transistor, and what is the maximum current this throttle can supply?  I have a slightly different circuit in mind (which will dissipate even less power).

Max DC power rating from MRC is 7 VA at 15 VDC, so call it 7/15 = 460 mA  (I'd say 400 mA would be more realistic because the output voltage really starts to drop off near 400).

The Darlington is a TIP 102G.  Looking at the data sheet, at around 1/2 amp, 4 volts, the gain curve shows a gain of about 2000.   The device is described at the top of the data sheet as having a "typical" gain of 2500.  So I'd say 2000 is a realistic number.  Bear in mind, my circuit modeler did not have the SPICE model for a TIP 102, so I had to just use a generic Darlington.