Copyright © 2006 Penton Media, Inc., All rights reserved.
Printing of this document is for personal use only.
Reprints

HI BOB,
Regarding quad op amps (“What’s All This ‘Free Amplifier’ Stuff, Anyhow?”), I thought I’d pass on this tidbit from my early days in the late 1970s. I was working at an industrial controls company on the east coast named Leeds & Northrup Co. (now defunct). (Yeah, I have collaborated with L&N. /rap)

At that time, they were producing a new line of industrial controllers with LED displays, rather than the older analog meter movements, to show process deviation from set-point. A stack of RC4136 “quad 741” op amps was used to drive the LED bargraph display. There were two display modes: single dot and bargraph. (That sounds like a Raytheon part number. /rap)

During burn-in, many of the bargraph displays would go into full-scale oscillation at a rate of about 2 to 3 seconds per cycle. Failure analysis showed that the current required for an RC4136 to drive four LEDs was enough to melt the VSS bond wire inside the package, but not enough to destroy the devI_Ce. The oscillation was the bond wire expanding, detaching, cooling, and re-contacting the bond pads.

(Wild! Was the amplifier in an extremely high-gain circuit? And, was it driving an excessive load? I know a 741 can typI_Cally drive a lot more than it’s rated to, but as you discovered, you can get in trouble doing this. /rap)

These devI_Ces would oscillate in this fashion during days of burn-in and still function just fine in single-dot mode when pulled out for final test! RAY BOWEN

HELLO RAY,
Wow, what a great story! RAP

BOB,
In looking over the schematI_C for the fancy PNP, I could not help but wonder about something (“What’s All This PNP Stuff, Anyhow?”). The current sources for sink and source are about 1 mA, while the current that is to be reflected into the output is 0.1 mA to 10 mA. Any mismatch in the sink, and source currents are going to show up in the output. (Yeah, but these currents are going to match well. That’s why I defined those current reflectors. /rap)

I might be missing something. If the base of the PNP is a low impedance point (Isn’t it a negative impedance point? So when current is dumped into it, it comes out the NPN’s collector and comes back out the PNP’s collector. /rap), you have a common base transistor configuration. And in this, the current gain is about 1. There is voltage gain, but the Miller C is removed as it is shunted to ground and does not reflect the driving circuit and cause slow down.
JIM P.

HELLO, JIM
You are a DAC man. Don’t be bamboozled by “paralysis by analysis.” It works! You could even build it in nine minutes.
RAP

SIR,
The Early effect is quite confusing to me. Most textbooks talk about the Early effect only in the CE configuration and do not give finer details. Why are we not observing the Early effect in the CB output characteristI_Cs? (Because it is beta times smaller in the grounded base, so it is just less notI_Ceable. Note: the Early effect in grounded-base is largely invariant of beta. In grounded-emitter, it is largely proportional to the beta. /rap)

Is it because we keep base emitter voltage constant to get constant I_E (No. If V_BE were kept fixed, the Z_OUT would get lousy again. /rap), unlike in CE where we need constant I_B? As in CE, I_C will increase due to a decrease in recombination at the base and due to an increase in V_BE (I_C = I_S eVBE/VT) but more due to the latter, and in CB, only recombination affects I_C. (Not really. You throw that around as if I_S were a constant. It is not. See what I said about IS in “What’s All This V_BE Stuff, Anyhow? Part 1,” at www.national.com/rap/Story/vbe.html, and in Part 2, June 21, 2007; ED Online 15690. /rap)

If so, then the change in I_C that we observe on the output characteristI_Cs shows the effect of base emitter voltage on collector current than that of base-width modulation. Is there a way to find the Early effect in CB configuration experimentally (measure hrb and measure hre; measure Z_OUT with fixed base drive, or fixed V_BE)? (Why is the output impedance at the collector about the same when I_B is held constant or when V_BE is held constant? Think about it. /rap)

I read somewhere that since the base is very thin when further reduction happens due to the Early effect, the injection of carriers from base to emitter also decreases. So, this also should decrease the base current. (I don’t think there is anything there to see. When you measure I_B, you can’t tell if it is from the injection or from the recombination. /rap) Exactly what things get changed due to the base-width modulation? (Many things. /rap) Please help.
GANESH NITHYANADAM

HELLO, GANESH,
Please, go measure some transistors. Help yourself. RAP

Comments invited! czar44@me.com —or:
R.A. Pease, 682 Miramar Avenue
San Francisco, Ca 94112-1232