Simple interview question

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
A few people have asked, so attached is one of my stock interview questions.

I have found that, from observing how people deal with this single question, I can usually develop a very firm opinion of their understanding of EE fundamentals and problem solving skills.
 

Attachments

paulktreg

Joined Jun 2, 2008
833
In your opinion it may be a simple interview question but to be honest my answer would be an educated guess. If you have a degree in electronics then maybe the answer is simple? Judging somebody on their ability to correctly answer the question, not sure about that one but it depends on the job in question.
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
What would the guess be?

Note that I said that the "question" was simple. As to whether it has a simple answer is a different matter.

I never really cared what answer they got, I wanted to observe how they approached the problem. I was certainly always aware that people are naturally a bit off their game in an interview situation.

I would first see what their immediate reaction and approach was. If they weren't headed in the right direction after a bit, either because they were completely confused or because they made an immediate but wrong assumption, I would as some leading questions intended to lead them toward an "ah hah" moment. I would continue in that vein taking note of how well they could apply fundamental concepts to the questions I asked and whether they had the ability to then apply the answers to those questions to the problem at hand.

I was also paying attention to what assumptions they made or didn't make and the degree to which their subsequent analysis reflected adherance to them.
 

#12

Joined Nov 30, 2010
18,224
It works for me. Defining my assumptions makes those circuits work (or not) and which assumptions I choose tells you how I think.
 

strantor

Joined Oct 3, 2010
6,782
I don't know what level detail is expected in the answer but from the level of detail and unknowns in the diagrams I would assume you're looking for a super simple answer. My simple answer would be that both circuits will output a 20Vpk-pk wave, with circuit B giving an inverted wave.
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
I have circuit b with a possible damaged B-E junction.
And of all the people that I ever asked this question of, very few ever got close to recognizing that, even after lots of hints (or worse). Only two (maybe even only one) saw that immediately. A couple of others found their way to it after just a few leading questions.
 

bretm

Joined Feb 6, 2012
152
My thought process was something like "hmm, first circuit seems like a simple follower with one diode drop of level shifting, I'll have to check more closely to make sure it....wait, WTH? Let's talk about circuit B first because that's a dead transistor. Maybe it was supposed to be PNP."
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
I don't know what level detail is expected in the answer but from the level of detail and unknowns in the diagrams I would assume you're looking for a super simple answer. My simple answer would be that both circuits will output a 20Vpk-pk wave, with circuit B giving an inverted wave.
For each circuit:

Q1) How do you arrive at a gain of 20?

Q2) What do you expect Vout to be when Vin is 0V?

Q3) What do you expect the base-emitter voltage to be when Vin is 0V?

Q4) What do you expect the emitter voltage to be when Vin is 0V?
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
My thought process was something like "hmm, first circuit seems like a simple follower with one diode drop of level shifting, I'll have to check more closely to make sure it....wait, WTH? Let's talk about circuit B first because that's a dead transistor. Maybe it was supposed to be PNP."
Pretty much the ideal response.

In trying to walk someone toward an epiphany moment, I kept simplifying things until I finally asked them what they expected to happen if I took a small silicon diode and connected it, forward biased, across a 12V car battery. They couldn't offer a reasonable mumbling.

I commonly got people throwing out the common-emitter and common-collector amplifier equations (and frequently not even at the correct circuit).

I frequently would remove the signal sources and just tie the input nodes to ground and ask what the base-emitter voltage should be, roughly, if the transistor is in the active region. Most (not all) could tell me something that indicated that it was a diode drop (which I would then say, let's assume it's simply 0.7V). So I would then ask what the emitter voltage is. This was much more problematical and perhaps half would come up with -0.7V (even for circuit (a)). Whatever they came up with for (a) they usually gave teh same answer for (b). I think one person actually said, for (b), -10V.

In almost all cases, I would end up at a contradiction where they were saying that the base-emitter voltage drop was 0.7V, making the emitter voltage -0.7V, and that the emitter voltage was -10V because it was tied to the supply. I was amazed by the number of people that didn't even seem to recognize that we had an obvious problem when we come to two very different voltages for the same node in a circuit.
 

strantor

Joined Oct 3, 2010
6,782
This is the part where I start shifting around uneasily and we both know where the interview is headed.
For each circuit:

Q1) How do you arrive at a gain of 20?
I didn't calculate any gains, I assumed saturated switches.

Q2) What do you expect Vout to be when Vin is 0V?
circuit A: -10V
circuit B: +10V
Q3) What do you expect the base-emitter voltage to be when Vin is 0V?
circuit A: 0v?
cricuit B: -10V? is this why people are saying it would blow up?
Q4) What do you expect the emitter voltage to be when Vin is 0V?
circuit A: -10v (yes, that means I think I was wrong initially)
circuit B: -10V

It took me >5min staring at the diagram to come up with these probably very wrong answers.
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
It works for me. Defining my assumptions makes those circuits work (or not) and which assumptions I choose tells you how I think.
Yes, though I tend to look closely at the assumptions involved in making (b) work, keepiing in mind that the assumptions are supposed to be reasonable.

Frankly, I never encountered anyone that offered any assumptions, even if not very reasonable, that would justify their answer for circuit (b). I was always hoping someone would say something about the internal resistances of the supplies or of the input signal source, to which I could then ask what they thought it would need to be in order for their analysis to hold up. Never got the chance. Had they gotten that far and pointed out that it depended on the transistor specifications, I had the datasheet for the small signal transistor we generally used on the lab ready to hand over to them. Never got the chance.

But while I expected people to have problems with (b), I expected people to get a reasonable answer for (a) without much assistance. I was very disappointed because very few people got that right (or even close), as well.
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
What kind of job would this interview be for?
For a position as an IC design engineer at a small company that does full-custom, mixed-signal ASIC designs. The person would be (eventually and rather quickly) doing circuit design, analog and digital simulation, digital logic design, layout, geometric verification, and database submission. The tools used were very low end. Separate layout editor, schematic capture, simulation, and design-rule checker. Nearly all schematics and layout hand drawn.
 

ErnieM

Joined Apr 24, 2011
8,377
I used to do some interviews. We had a standard "test" of 10 questions we gave candidates, same test to EE's and techs. A tech got extra credit for solving anything, but what I was looking for was him to follow my explanation and be able to feed some of it back to me.

EE's had to solve the damn things. Most couldn't, and they were not tricky questions. THIS example is a tricky question.

One interview I was given an on the spot sketch of 3 voltage sources and 3 resistors asking the voltage between 2 points. Not liking hard problems I did a thevenin equivalent to combine two of the sources so the answer would be obvious. The guy asking me was quite impressed that I actually remembered what a thevenin equivalent was, and even though he told me I got the wrong answer (which he quickly proved by some magical technique) I got the job.

I sketched the circuit later and gave it to a sim program. It agreed with my answer, not his.
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,978
I agree that it can be regarded as a tricky question, though almost any EE course that is introducing students to BJTs is going to deal with the issues in (b) pretty thoroughly (I hope).

I would NOT give this question to someone that was answering it all alone and then turning it in for grading; the whole point was to see how they dealt with a seemingly simple circuit that had a not-too-obvious complication. Could they identify that a problem even existed (most could not), could they articulate what the essence of the problem was (very few could even come close), could they describe a likely outcome if the circuit where actually implemented (and in one round I actually had a solderless breadboard, two 9V batteries w/clips, a 10kΩ resistor, and some hookup wire for them to wire up the grounded-input version of the circuit -- which most could not do).
 

bretm

Joined Feb 6, 2012
152
I didn't calculate any gains, I assumed saturated switches.
After assuming the operating mode and determining the consequencies, you need to verify the assumption. In this case, an assumption of saturation means a very small Vce and as a rule of thumb, a base current greater than about 1/10th the collector current.

If the transistor is in saturation, the Vout of circuit (a) would be near +10V, which means the base-emitter junction would be reverse-biased (since Vin never exceeds +1V). So the assumption can't be right.

The transistor in circuit (b) is in trouble because the base is between 9V and 11V above the emitter, so the base-emitter junction is very forward-biased and is essentially short-circuited. Large current * large voltage = large power = large heat.
 

bretm

Joined Feb 6, 2012
152
....for them to wire up the grounded-input version of the circuit -- which most could not do).
Wait...what? They're applying for a circuit design job and they can't breadboard a circuit with one transistor and one resistor? I guess after the MIT lightbulb video I shouldn't be surprised, but I'm still amazed by that. But you say most could not do it. Not just not analyze it, but actually not be able to wire it together and see what happens. I don't get it. That seems like a huge breakdown in the teaching/learning process.

I don't think it's a trick question at all in the context of a job interview. It's an easy mistake to put an NPN into a schematic where you meant to put a PNP. I'm in software, not hardware, and a "what would this code do" type of question is very much fair game.
 

steveb

Joined Jul 3, 2008
2,436
I always ask a EE if they know Maxwell's equations. I was taught that every EE should know them by heart. But, in my experience, too few actually know them. :(

I give an A to anyone who knows them and a B to anyone who has them written or tattooed on their arm. I'll give a C to anyone who is embarrassed that they forgot them. But, the ones that act like I'm crazy for thinking a EE should know them get an F in my book.

Keep in mind that this standard does not even scratch the surface of whether they are able to understand and apply them, and even fewer can do this.
 
Top