Re: High Side Pulse Transmitter

On Fri, 14 Oct 2005 02:23:44 GMT, Fred Bloggs <nospam@xxxxxxxxxx>
wrote:

>
>
>Winfield Hill wrote:


>>
>> Right. Judged by the O.P.'s specs, that one is a bunch of
>> conceptual garbage. Also, it doesn't meet most of the specs,
>> e.g. it's inverting, etc. I'm sure it's not John's circuit.
>
>I think he was planning on use complementary MOSFETs.
>
>> Certainly it's not what I had in mind.
>>
>> Here are Jeff Stout's specs:
>>
>> 1. True logic; high input produces high output, etc.
>> 2. Operates with Vcc from 24V to 50V.
>> 3. High output is slightly less than Vcc, low is better than
>> 6V less than Vcc (exa, Vh = Vcc - 0.5, Vl = Vcc - 6.0)
>> 4. Rising and falling edges should be as fast as possible.
>> 5. Load is from 10mA to a maximum of 125mA, and capacitance
>> from 0 to about 1.5uF.
>> 6. spends most of its time in the high state.
>> 7. It transmits a signal between 1200 BAUD, and 9600 BAUD
>
>>
>> Here's something like the five-part circuit I imagine John had
>> in mind. It can provide a non-inverting 6V output with better
>> than 10us risetime, assuming high-gain transistors, like those
>> sold by Zetex. The circuit features a level-shifting current
>> source (sink actually), which also sets the 6V output swing for
>> the emitter followers. A power resistor can be added in Q3's
>> collector for current limiting (the 6th part John mentioned).
>>
>> . ------+------+---- + rail
>> . | | 24 to 50V
>> . 1.0k |
>> . | |/
>> . +----|
>> . | |\e
>> . | Q3 | noninverting 6V swing
>> . | +--- output to 1.5 uF load
>> . | Q2 |
>> . | |/e Imax = 750mA
>> . +----| BJT beta > 150
>> . | |\ dv/dt = i/C = 6V in < 12us
>> . Q1 | |
>> . |/ |
>> . +5 ----| gnd etc
>> . |\e |
>> . 620 | V 6.3 mA for
>> . o--/\/\---' logic LO
>> . IN: 5V cmos logic
>> . with 60-ohm gate Ron
>>
>> Just to show it can be done with five parts. A few more parts
>> could be used to improve the circuit, of course. Jeff can test
>> it with spice, if he has the right transistor models.
>>
>>
>
>Now you know that circuit is unusable- so you haven't "done" anything
>with those five parts.


What's unusable about it? It's simple, it satisfies all the stated
requirements and, more important, it should work in the application.

John

.

Quantcast