# Discrete 741 build final project

#### SamR

Joined Mar 19, 2019
4,303
Those and some male to male 90-degree header pins do the trick for breadboarding.

#### dl324

Joined Mar 30, 2015
14,710

#### SamR

Joined Mar 19, 2019
4,303
@dl324 Not sure what I am looking at Dennis? Link and maybe a description of how they connect to the adaptor boards?

#### dl324

Joined Mar 30, 2015
14,710
They're called leadframes. I have some made by Tyco, Die-Tech, and Proner Comatel.

The ones used here have 2 prongs. The previous picture was for a 3 prong variety.

When you can find them, they sometimes cost more than the adapter. In this case, 3 pins cost $0.30 and the adapter costs$0.04.

They're used in things like resistor and other networks. Price comes down if you buy full reels of 30,000 or so.

Last edited:

#### Wolframore

Joined Jan 21, 2019
2,554
So before I committed, I checked hFE using a DMM and glad I did... now all the beta and Vbe matches for the current mirror as well as the differential inputs.

Now I'm stuck... I might have to run out and get some 9 volt batteries, I usually work on digital so I've gotten by with the single DC supply. I'm going to have to look into a dual supply.

#### dl324

Joined Mar 30, 2015
14,710
I usually work on digital so I've gotten by with the single DC supply. I'm going to have to look into a dual supply.
If you have single supplies with isolated ground, you can stack them and use the common connection as circuit ground.

#### Wolframore

Joined Jan 21, 2019
2,554
If you have single supplies with isolated ground, you can stack them and use the common connection as circuit ground.
Aha, thanks, I just found two 12V wall warts and the grounds are isolated! I have a bunch of errands to run today, time to plug in my sig gen.

#### cmartinez

Joined Jan 17, 2007
7,750
Aha, thanks, I just found two 12V wall warts and the grounds are isolated! I have a bunch of errands to run today, time to plug in my sig gen.
Wall warts are very noisy sources of voltage... Trust me, I've had my fair share of headaches because of them. But what I've done more or less successfully in the past is install an inductor (about 50 µH) and a fat cap (in the order of 4,700 µF) one after the other right at its output. That small addition makes a ton of a difference, noise-wise.

#### Wolframore

Joined Jan 21, 2019
2,554
LC LPF got it.. RF choke with bypass, I will see what I have on hand.

Last edited:

#### dl324

Joined Mar 30, 2015
14,710
I have a Heathkit experimenter that has a dual supply. The power supply circuit is really weak and it has a couple hundred millivolts of ripple. It rarely affects any of my circuits; though it did have me scratching my head on a couple circuits where it prevented correct operation.

One of these days I'll improve the power supply. That is if I ever find myself without something better to do.

#### schmitt trigger

Joined Jul 12, 2010
497
@Wolframore:

I wish I had found this thread earlier, to be able to give you some hints. But many here already have done so, and your finished project is simply awesome!I hope you get the full grade you deserve.

On another subject: I read somewhere that the classic Bob Pease photo is of him "airboarding", is actually the prototype for the LM331 voltage to frequency converter.

#### Wolframore

Joined Jan 21, 2019
2,554
I just tested the integrator against an actual uA741. I finally got rid of the grandkids so I can actually work. It works! Small clipping but the power supply is a little off... I didn’t do any offsetting. The test equipment is subpar. Power supply was quiet enough without extra filtering.

actual 741 shows a little clipping also.

Results from my discrete 741

#### Wolframore

Joined Jan 21, 2019
2,554
I’ve damaged the board shorting the resistors. Do not short R5. It’s burnt out Q11... so Q10 and Q11 will be replaced to keep the current mirror matched although since it’s a Widlar mirror it doesn’t matter as much. It damaged the BE junction. It now drops over 2v... glad I found it quickly.

yep it fixed it. I did notice that the artifact reduced quite a bit when R10 is shorted. I should make a jumper with a resistor it would be safer

Last edited:

#### cmartinez

Joined Jan 17, 2007
7,750
I’ve damaged the board shorting the resistors. Do not short R5. It’s burnt out Q11... so Q10 and Q11 will be replaced to keep the current mirror matched although since it’s a Widlar mirror it doesn’t matter as much. It damaged the BE junction. It now drops over 2v... glad I found it quickly.
Ouch!

#### Wolframore

Joined Jan 21, 2019
2,554
Fixed it and fixed the artifact. It just needed a pot on the offset!

playing with the screen capture... it will be nicer for the presentation. Jut noticed the time said it was past Turkey day and I haven't even had my turkey yet... The power supply is noisy but doesn't matter for this op-amp, it poops out at 10 kHz so it won't matter much.

Last edited:

#### cmartinez

Joined Jan 17, 2007
7,750
Fixed it and fixed the artifact. It just needed a pot on the offset!

View attachment 193332

playing with the screen capture... it will be nicer for the presentation. Jut noticed the time said it was past Turkey day and I haven't even had my turkey yet... The power supply is noisy but doesn't matter for this op-amp, it poops out at 10 kHz so it won't matter much.

View attachment 193334
Hey, I think I found the problem. That's a soda can, right? ... had it been a beer can I'm certain you would've got it right the first time...

#### Wolframore

Joined Jan 21, 2019
2,554
Success - after burning out 3 boards - including the traces on 2... I've finally succeeded in a closed loop test (without smoke). It takes about 2-3 hours to match transistors and solder another board.

One off the issues was trying to do a close loop without resistors. It was also overloading the signal generator and would not turn on once everything was correct. Using 100k at the fb and 100k on the signal input solved it (with smaller signal in).

It starts to drop off after 100 kHz with gain of 1. The interesting thing is that it's passing the signal at 9 MHz... it's lost gain at this point but still visible. I need a better oscilloscope but again there's some oscillation but it's a uA741!

Some of the issues as pointed by by other member and myself, there are a few different schematics floating with different resistor values... I've finally found the correct values for them from trial and error, by maxing out the values to protect the circuitry, then reducing them based on the different schematics to where it is now.

I will post a follow up with the correct resistor values. The professor thinks I should contact Lady Ada and offer the kit online

#### Wolframore

Joined Jan 21, 2019
2,554
Update: Class is over - presented to class my grade was 95

I discovered another issue the morning of the final class: this is a dual collector transistor - something that can be built inside a chip but not available as discrete.

Due to the splitter from the current mirror, I was unable to get good VBE generator to the output stages. Also there are so many schematics floating around, I ended up burning out quite a few transistors until I was able to find the correct values.

Solution was to solder another transistor on top of Q13 and cut the trace and place a wire to the VBE generator from the new transistor. It is working now. Here's the set up for closed loop gain of 1.

and the results: 6.7 MHz to the -3dB point I can go to 2MHz @ Av=1.. LTSpice was correct, my discrete opamp has much higher bandwidth than the IC version. One issue to work out, I have a 600mV positive offset... not sure what's causing it yet.

#### ci139

Joined Jul 11, 2016
1,864
One issue to work out, I have a 600mV positive offset... not sure what's causing it yet.
the offset can be internally adjusted from R3 in "uA741 - MDev - 2h.asc" (at next post) ← but it seems to be supply and frequency dependent
or by compensation pins . . .
__________
a quick compile (looks like the R4 sets the upper bandwidth . . .) haven't tested for stability overshoot ringing offset ....
some improvements + a load of tests (there's yet no common mode range & sensitivity test + more)
just for comparison
the thing seems a bit unstable (the actual µA741 has supposedly greater parasitic series base resistances)

Last edited: