Voltage Comparator Circuit Needed LM311 Equivalent

[amdkicksass]

I am building a circuit that incorporates a lm311 as a voltage comparator. I need to replace the LM311 IC with a circuit that is completely discrete. By discrete i mean, resistors, fets, transistors, diodes...etc.
I have seen a few circuits on the web, but they incorporate over 20 transistors, which might make things a bit difficult. IT doesn't have to be perfect but work decently in my successive approximation a/d circuit and i think i am making thing more difficult than they should be.

I am guessing a circuit thats based on a basic differential amp would do the trick but i am not sure.

Thanks
Pete

 

[HarveyH42]

Ir's only a guess, but if there was a small, simple way to do it, there wouldn't have been a need for the IC. Usually the Data Sheet shows an equivalent circuit, or at least a block diagram of what's going on inside the chip. Might give you an idea of what you need.

 

[amdkicksass]

I tried looking at the lm311 datasheet and the diagrams they show are very theoretical and probably impractical for this project.

I wonder if this can be done using any 74 series chips ?

pete

 

[Papabravo]

In short no. The 74 series are logic devices which respond only to low and high levels. A comparator is an analog circuit, usually with an open collector output to interface to a logic circuit.

Why on this great grreen Earth would you want to build a discrete comparator? Is there some reason why the LM311 is unsuitable for your design?

 

[amdkicksass]

Well The project is to build a A/d converter. I have already pretty much built the converter with the lm311 IC for testing purposes, but one of the specs/rules is that no ICs can be used and only discrete components. The only exception is the 74 series digital ICs may be used.
So in short i am now looking to replace the comparator with a simple discrete circuit.
You may refer to the A/D threads for more info..
https://www.electro-tech-online.com...tal-converter.28993/?highlight=analog+digital

THanks
Pete

 

[Gayan Soyza]

Here is LM311 with discrete components of course you need lot of transistors.


**broken link removed**
Here is another comparator circuit’s site which may useful to you. It was helped me a lot.
So I like to share with you.

**broken link removed**

 

[amdkicksass]

I think maybe i should rephrase my questions differently.
Is their a simpler circuit that can possibly fulfill the same requirements in this particular circuit (A/D Comparator) without being too complicated. The circuit doesn't have to have all the functions and properties of a lm311 but rather do its job in this situation.

pete

 

[3v0]

What range of voltages does you ADC need to measure directly?

 

[Optikon]

I think maybe i should rephrase my questions differently.
Is their a simpler circuit that can possibly fulfill the same requirements in this particular circuit (A/D Comparator) without being too complicated. The circuit doesn't have to have all the functions and properties of a lm311 but rather do its job in this situation.

pete

Look at datasheet for LM393.. its a comparator with not too many transistors.

**broken link removed**

Doing a discrete A/D converter that works well is no easy task. Your professor is being silly by making you build comparators out of transistors.

 

[Roff]

I posted a comparator circuit for you here. You never said what was wrong with it. It has 11 transistors. You need lots of trannies because you need high gain, high speed, low offset, good temperature stability, etc.
I simulated that circuit before I posted it, and it should work for you. I'm frankly pissed that you are asking the same question again without as much as a thanks, go-to-hell, or anything.

Ron

 

[Papabravo]

I think being pissed is the mildest of reactions. I congratulate you, Roff, on your fortitude. I would be positively apoplectic if I had gone to such lengths without so much as a by your leave.

 

[pc88]

Well The project is to build a A/d converter. I have already pretty much built the converter with the lm311 IC for testing purposes, but one of the specs/rules is that no ICs can be used and only discrete components. The only exception is the 74 series digital ICs may be used.
So in short i am now looking to replace the comparator with a simple discrete circuit.
You may refer to the A/D threads for more info..
https://www.electro-tech-online.com...tal-converter.28993/?highlight=analog+digital

THanks
Pete

Build a voltage to frequency converter. Will that satisfy the requirements for your assignment?

 

[HarveyH42]

I wonder why a school would waste time on such a thing. I can't imagine any practical reason to do this with transistors in these modern times. 30 years ago, it would have been a useful challenge, but now there are single chip solutions, and its a chip feature on many microcontrollers. The LM311 is an old chip.

 

[amdkicksass]

First of all i would like to apologize to Roff, for any wrong impressions i may have given to you. I indeed do appreciate your efforts with that circuit.

My purpose in making this post was to get more exposure and ideas for that one issue (comparator) . I am sure many will agree that there are many ways to doing the same thing. My goal was to get other ckts that may be simpler or rather fulfill the criteria differently.

Once again Roff as well as other, i have appreciated your help along this project and will continue to do so.

Harvey i am unsure why the professor would want us to do this for the comparator. Especially when the goal is a a/d converter not a comparator. But i can honestly say tho building this project, the a/d converter has given me valuable insight to how they operate and general knowledge.

PC88, the comparator is being used as part of the a/d circuit which as far as i know has nothing do with frequency, so i don't think that would help in the a/d process. Thanks for the suggestion tho..

Pete

 

[pc88]

PC88, the comparator is being used as part of the a/d circuit which as far as i know has nothing do with frequency, so i don't think that would help in the a/d process. Thanks for the suggestion tho..

Pete

I was just suggesting a different way to do the A/D conversion. If you can convert a voltage to a frequency or duty cycle, digital logic units can complete the conversion to a digital value.

 

[Roff]

I was just suggesting a different way to do the A/D conversion. If you can convert a voltage to a frequency or duty cycle, digital logic units can complete the conversion to a digital value.

He needs 100kHz signal bandwidth. I think this would be extremely difficult to do with a V-F converter.

 

[amdkicksass]

Thanks for the suggestion but unfortunately it has to be of certain type, namely the successive approximation type- Therefore i don't think frequency conversion would help. Good thought tho..

Hey Roff whats with the RF transistors, can one use standard 2n2222a and hobby type PNP transistors or do i have to use those listed on the schematic ?

Thanks loads..
pete

 

[Roff]

Thanks for the suggestion but unfortunately it has to be of certain type, namely the successive approximation type- Therefore i don't think frequency conversion would help. Good thought tho..

Hey Roff whats with the RF transistors, can one use standard 2n2222a and hobby type PNP transistors or do i have to use those listed on the schematic ?

Thanks loads..
pete

I tried substituting 2N3904 for all the NPNs, and 2N3906 for all the PNPs, and then rerunning the sims. It might still work at 100kHz signal bandwidth, but I doubt it. The propagation delay is way too long. You need delay times of considerably less than a microsecond, because your clock rate is around 1MHz. The DAC and the comparator combined have to settle in less than a clock cycle. I also doubt that your 10k R2R ladder will settle fast enough.

 

[Optikon]

You guys must have a different thread going on this, I dont see where he said he needed 100k signal bandwidth nor any circuit showing RF transistors. Confusing as to what the current state of affairs is..

 

[Roff]

You guys must have a different thread going on this, I dont see where he said he needed 100k signal bandwidth nor any circuit showing RF transistors. Confusing as to what the current state of affairs is..

Amdkickass should not have started a new thread. the original one is here.