Level Shifter

[bagydor]

Hi, Don't worry. I searched the forums before asking this Found some examples of how to do the opposite of what I want to do, but nothing that I could transmogrify to my needs

I need to make a level shifter which will change TTL HI/LO (about 0.8V and 2.4V I think), values to +15V and -15V accordingly. I've tried simulating some things but I just can't get it to work I'm not asking people to do this for me (I would like to complete it myself), but I think I'm stuck, so if ye could give me pointers on how I may go about doing this or links to additional resources online which will aid me I would be very happy! I obviously have +/-15V supplies and can only use transistors/resistors/capacitors.

 

[Nigel Goodwin]

First thing you need to do is tell us what you're trying to do, the description is too vague otherwise.

 

[bagydor]

The +/-15V is going to go into the gates of two CMOS transistors. It's for a switch, so I have a 0V -> 5V input being fed into an inverter (which I have working), giving me a 5V -> 0V output of that stage. The switch itself consists of two CMOS transistors (one P and N type in parallel), which are controlled by the gate voltage, which has to be +/- 15V. So it's the part between that I must get working.

Need more?

 

[Nigel Goodwin]

Doesn't really make any sense to be honest?, all you're wanting is to invert the 5V logic signal? - just use two resistors and an NPN transistor.

 

[Roff]

The +/-15V is going to go into the gates of two CMOS transistors. It's for a switch, so I have a 0V -> 5V input being fed into an inverter (which I have working), giving me a 5V -> 0V output of that stage. The switch itself consists of two CMOS transistors (one P and N type in parallel), which are controlled by the gate voltage, which has to be +/- 15V. So it's the part between that I must get working.

Need more?

A CMOS transmission gate requires complementary signals to the two gates, i.e., +/-15V on the N-channel, and -/+15V on the P-channel. Is that what you need? If so, do you mean that you are going to take your input signal, and the inverse of it (which you said you have) and drive a level shifter with each of them?
To make a level shifter, think of an inverter that translates 0 -> 5V to 0 -> 15V. From that output, think about how you can make another inverter that swings from +15V -> -15V. Remember you can use PNPs as well as NPNs (on PMOS as well as NMOS).

 

[bagydor]

>> Doesn't really make any sense to be honest?, all you're wanting is to invert the 5V logic signal? - just use two resistors and an NPN transistor.

Hmm. I'll attach an image of it ... God knows I'm not gonna write a thousand words



That make more sense? Just did it on paint ... so bear with it. I have everything besides my black box designed and simulated. It is possible that I have misinterpreted the assignment .... but I don't think I have.


EDIT: Yeah. That's what I mean, Ron. I'll consider what you suggested and come back here if I can't think of a way to do that, or if I do and am struggling to make it work.

 

[Nigel Goodwin]

What EXACTLY does your assignment say?.

 

[ericgibbs]

hi,

If you just want to level shift TTL to +/15v and you are not bothered about signal inversion, use a MAX232 with +/-15v supply rails.

If inversion is a problem,use a TTL invertor to drive the MAX232.

If required, the MAX232 will voltage level 'shift' a +/-15v signal to TTL levels

A nuts and sledge hammer solution.

EricG

 

[hjames]

What a confusing request. If you start out with .8V/2.4V TTL (i.e. BJT) style logic, why bother messing with transmission gates? A common base NPN and a pullup resistor is all you need. I guess a common gate mosfet would work too.

Obvious hint: go to the Texas Instruments website and look at their logic level translators. Find a data sheet, read it, and steal all the good bits.

 

[Roff]

What a confusing request. If you start out with .8V/2.4V TTL (i.e. BJT) style logic, why bother messing with transmission gates? A common base NPN and a pullup resistor is all you need. I guess a common gate mosfet would work too.

Obvious hint: go to the Texas Instruments website and look at their logic level translators. Find a data sheet, read it, and steal all the good bits.

My understanding is that the transmission gate is a given. Maybe he has a high-level analog signal he wants to switch. His problem is getting from 5V logic to +/-15V.

 

[bagydor]

What Ron H said. I probably wasn't very exact in my explanations. My bad. In my paint-diagram above I made a mistake ... on the switch side (RHS), the Vin should be Vin2 ... or the LHS Vin should have some sort of controlling reference in its name.

Think about it as an analogue switch which can be turned off by TTL signals. LO => Switch off (open circuit), HI => switch on (closed circuit). That make more sense? But it (the switch), must handle +/- 15V which is why I need the shifting in the first place.

However, I've come up with a solution to my problem, thankfully, somewhat of an abstraction of Ron H's hinting above. Queue paint ... Nah. Only kidding. Have a p-spice simulation to screen shot:



The idea is that I subtract 2.5 volts from the output of the inverter, which will either give me 2.5V or -2.5V depending on the output. I then amplify that by 6 and -6 which will give me my desired gate levels. I don't yet know how to do this so I used the magical components that p-spice let me use, sum and mult blocks in this case. Could someone throw some pop words about referring to things I should go research to do the summing and multiplication blocks? Obvious answer: search Google. Have done, but I'm getting a lot of holiday advertisements ... yes. Even with the words transistor

Obviously, what ericgibbs suggested would be perfect but the problem about that is that it's a design assignment and I think we are meant to use transistors and as little silicon as possible ... hjames ... that not called plagiarism

NOTE: Attachment is in place of the black box posted yesterday.

 

[Roff]

An analog solution to a digital (binary) problem is not optimal. From your replies, I gather that you don't understand how to make switching circuits from bipolar transistors - or MOSFETs either, except in CMOS form, like the inverter you drew. Why have you been given an assignment for which you are so ill-prepared? I could draw you a schematic, but mostly you would just learn that you can get your problems solved by going to a forum.

 

[ericgibbs]

hi,
As Ron says it would take 2 minutes to sketch out a PNP transistor solution, to give the OP the explicit answer, BUT in the long term that wouldn't help the OP to LEARN how to do for himself.

Thats why his tutor asked the question in the first place.

The tutor I 'hope' is looking for the 'way' the OP got the answer.

EricG

 

[bagydor]

Sigh.

Contrary to the impression I seem to have passed onto ye guys, I am not an idiot. Nor am I ill prepared in my assignment. I am not a genius at circuit design, as I think ye seem to have already grasped, but that doesn't mean I'm not trying. Where in my previous replies have I asked for a solution? This is not the only forum of which I am a member. I know people don't like to give solutions to problems. I know people learn best when they try things themselves. That's why I've been trying.

>> I could draw you a schematic [...]
>> As Ron says it would take 2 minutes to sketch out a PNP transistor solution, to give the OP the explicit answer, BUT in the long term that wouldn't help the OP to LEARN how to do for himself.

Please tell me where I asked for a schematic ...

>> if ye could give me pointers

I thought from the outset I made it clear that I wanted to do this assignment myself.

Now don't get me wrong - I really appreciate the help ye have given me to date but I just need to make clear to ye guys that I believe ye are misjudging me.



Ok. So are ye suggesting I make a BJT summer and amplification stages to get my desired voltage levels? That much I can do, and I know how to do. Or would Ron H's three-inverter solution work better?

I'm new to designing circuits, that much I think is obvious, but I really want to learn! We haven't had many design exercises so far. I'm green. I don't know the tricks that all of ye probably know better than the back of your hands. I'm damn eager to learn if truth be told. Sorry if I came across as abrupt and if I seemed like I was demanding solutions.

 

[Roff]

Please tell me where I asked for a schematic ...

OK, I'll throw it back in your face: Who said they thought you were an idiot? On the contrary, I can tell you are intelligent, but do not have the background to solve this problem. Do you want me to post a schematic? Perhaps just some more clues? I'm not sure what I can say short of giving you a working circuit.

 

[bagydor]

>> Who said they thought you were an idiot?
Nobody. Sorry.

>> but do not have the background to solve this problem.
I agree.

>> Do you want me to post a schematic?
I would like to say yes, but no thank you. Not yet anyway. I'm not just saying this, I really believe it - people do learn best when they have to figure problems out. I've seen it happen, so I think it's best if I go away, try to make solve the problem maybe post my attempts if I fail.

Just one question before I do that - Would you recommend I do what you suggested earlier (0->5; 0->15; -15->+15), or what I suggested? And if it's the former would a BJT solution be best? Or MOSFET/CMOS? Thanks.

 

[Nigel Goodwin]

I thought from the outset I made it clear that I wanted to do this assignment myself.

I asked earlier, but you never responded, what was the actual assignment?.

 

[Roff]

Do you have any switching speed requirements? That may have some bearing on the method you use.

 

[bagydor]

Sorry, Nigel. Is, my good man. Is. Present. Ye familiar with the **broken link removed**? It's basically that but with some differences ... ton = toff = 100 ns .... Ron = 25hm: and a bandwidth of 100 MHz, Is(off) = 1 nA. All those parameters I can deal with in due course, but first off I want to get the layout right before I start tweaking it, if ye get me.

 

[Roff]

Sorry, Nigel. Is, my good man. Is. Present. Ye familiar with the **broken link removed**? It's basically that but with some differences ... ton = toff = 100 ns .... Ron = 25hm: and a bandwidth of 100 MHz, Is(off) = 1 nA. All those parameters I can deal with in due course, but first off I want to get the layout right before I start tweaking it, if ye get me.

Are you using discrete MOSFETs for the analog switch? If so, what are the part numbers? This matters a lot in the design of a driver.
EDIT: ADG201A isn't an assignment, it's a part. What are you doing with the analog switch?