[GreenKeys] Snubber tests

[Bob Camp]

Hi


One thing I should have mentioned. Hooking up the magnets for 60 ma  
operation is a better idea than 20 ma operation. That alone cuts the  
voltage on the switch in half. It's been about (gulp) 40 years since  
I played with a 2B so I don't remember all the details. I do believe  
that the magnets are "normal" and can be set up for 60 ma.

As you have found running straight off of a high voltage opto is not  
even a real good thing at 20 ma. To saturate the switch and keep it's  
dissipation low you need to significantly overdrive it. The same  
thing applies to a normal bipolar transistor. If you have a gain of  
200 you drive it like it had a gain of 40 or so.

Bob

On Apr 12, 2007, at 12:31 AM, Peter Gottlieb wrote:

> Aaack!  It's late!  That was a 10K, not 100K resistor!!!
>
>
> Peter Gottlieb wrote:
>> I just ran a few tests.  The simplest snubber which still  
>> absolutely protects the MJE13005 consists of a diode (I use  
>> 1N4007's for just about everything) in series with a 100K 2W  
>> carbon resistor across the selector magnet.  The voltage on  
>> opening goes up to around 500 volts then shows a nice exponential  
>> decay.  The resistor gets mildly warm after steady copy, probably  
>> a 1 watt would do fine.
>>
>> So now I hooked up an optocoupler as the input from the low level  
>> data line.  I first tried a TIL111 because I have a bunch of them,  
>> no good, not enough current transfer.  The best I could find I  
>> scrounged from an old switching power supply board and its  
>> datasheet says it has a 200% CTR (Current Transfer Ratio) yet I am  
>> not seeing anywhere near that and it is marginal.  I must be  
>> missing something.  I am driving at 20 mA and only need 17 mA on  
>> the output.  I think I am expecting it should go into saturation  
>> and it isn't designed to do that.  I may need additional  
>> components, maybe make it like a darlington or use a tiny FET (I  
>> used to use the 2N7000 a lot, must have some still around somewhere).
>>
>> In any case, I did get data from the Dovetron through to the 2B.   
>> There was no QRM generated at all, at least not that I could  
>> detect even on very weak signals.  However, I was unable to get  
>> good copy on any signals.  One problem was that many signals were  
>> not 22 msec timing (ie, not 45.45 bauds).  One that was, still no  
>> joy, and I note the 2B is not spacing evenly.  It sometimes  
>> doesn't space properly, or at all, when cold so I have to get into  
>> that part of the mechanism and lubricate, clean and adjust it as  
>> necessary.
>> Getting there.  Once the opto problem is resolved I need to put  
>> one in for the other direction so the keyboard runs the Dovetron.   
>> Then add a full/half duplex switch (hmmm, I wonder if the Dovetron  
>> switch does that?) and build the final loop supply/signal  
>> converter into a suitably boatanchory looking box.  I did note  
>> that the Dovetron "standby" switch does nicely enable/disable the  
>> printing.
>>
>> Peter
>>
>> _______________________________________________
>> GreenKeys mailing list
>> GreenKeys at mailman.qth.net
>> http://mailman.qth.net/mailman/listinfo/greenkeys
>>
> _______________________________________________
> GreenKeys mailing list
> GreenKeys at mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/greenkeys
>