Electronics > Projects, Designs, and Technical Stuff
Voltage comparator
<< < (3/8) > >>
Benta:

--- Quote from: OM222O on April 29, 2019, 03:44:28 pm ---
--- Quote from: Benta on April 29, 2019, 03:27:27 pm ---The classic way as done by HP and others is a JFET input stage followed by an MC10H115 or MC10H116 line receiver with the stages in series.


--- End quote ---
can you please post a schematic of that? It would be very helpful thanks!

--- End quote ---

Unfortunately I only have these crappy screenshots. The schematic is equivalent to the input stage of an HP 20 MHz counter.
A1 and A2 are 1/2 MC10H115, A3 is a fast comparator.
It's a nice design, input sensitivity is 50 mV up to 5 V, with attenuator up to 100 V and not really sensitive to overload.

IDEngineer:

--- Quote from: David Hess on April 29, 2019, 05:24:13 pm ---Operational amplifiers can do all kinds of weird things when driven into saturation.
--- End quote ---
x2 on that statement, which I didn't say myself because I thought it was common knowledge. Can you use an opamp as a linear comparator? Of course. Should you? No. There's a reason they make dedicated linear comparators, and it's not just speed.

Again, if you can keep this in the digital domain with a Schmitt Trigger that would be an excellent solution.
OM222O:

--- Quote from: David Hess on April 29, 2019, 05:24:13 pm ---At 3 microseconds for a TLV7031, a 20 MHz input pulse could and would likely be completely missed.
TI does make suitably fast parts; perhaps the LMV7219 would be suitable.

--- End quote ---

As far as I can tell from reading data sheets and a quick google search:
propagation delay is just what the name implies: a delay between the output and input where output is supposed to follow the input. by that logic no matter how fast the input is, the output will be the same, but with x amount of delay  :-// rise and fall times are what limit the output speed as far as I can tell. please correct me if I'm wrong.
OM222O:

--- Quote from: IDEngineer on April 29, 2019, 05:30:32 pm ---
--- Quote from: David Hess on April 29, 2019, 05:24:13 pm ---Operational amplifiers can do all kinds of weird things when driven into saturation.
--- End quote ---
x2 on that statement, which I didn't say myself because I thought it was common knowledge. Can you use an opamp as a linear comparator? Of course. Should you? No. There's a reason they make dedicated linear comparators, and it's not just speed.

Again, if you can keep this in the digital domain with a Schmitt Trigger that would be an excellent solution.

--- End quote ---

as I mentioned that's simply not possible given the application. what's the best way of implementing a comparator as an schmitt trigger with variable voltage reference (again, switched by fets)
The entire reason I'm bringing this up is I actually saw the chinese clones of the frequency counter, which as you said use the PIC uCs to create a very simple frequency counter, however the input is directly connected to the pin of the MCU which is not very good. the original designer has suggested using a pre amplifier to boost the input using a BJT. I just wanted to do it with a digital circuit rather than a common emitter amplifier because as you said, my particular one does not include schmitt triggers built on the IO pins. I'm not sure if the ICs like those slow rising edges of sine waves for example (at low frequencies that is). so I'm open to suggestions as to how I can solve this issue.

Edit:
here is the original article: https://www.qsl.net/dl4yhf/freq_counter/freq_counter.html
many people have tested it and it works surprisingly well (the 5 digit version is the one that chinese have cloned).
David Hess:

--- Quote from: OM222O on April 29, 2019, 05:27:00 pm ---I would really like to know how name brand manufacturers like HP did it as Benta mentioned. I really wish he had provided some sort of schematic , more details or where to find those details ...
--- End quote ---

I can tell you how they did it and it was not by using operational amplifiers as comparators.

If the input impedance was high, then first an impedance buffer was used like you would find in an oscilloscope.  Then a cascade of limiting amplifiers was used followed by a line receiver, logic gate, or sometimes a two transistor comparator with hysteresis.  A modern integrated comparator can replace the cascade of limiting amplifiers to drive logic directly in lower frequency (below 100 MHz) applications.

Check out the various old Tektronix and HP frequency and universal counter designs which are fully documented.

The example below is from the 125 MHz Tektronix DC503A which does the high impedance conversion using a dual FET to drive a chain of 10216 ECL line receivers used as differential limiting amplifiers.  The output from the amplifier chain just goes into a normal ECL gate.  A discrete implementation would add hysteresis to the last amplifier stage.  Everything shown except the input FET can be replaced by an LT1016 or similar. (1)

I have a modern (20 year old) counter which uses an LT1016 for its 10 and perhaps 50 MHz inputs.

(1) Do not bother looking for a fast comparator with high impedance inputs to avoid having to use a separate FET input buffer.  Such a device is a contradiction in terms.

Update: I am not sure what happened to the example so I added it back.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod