What a circuit diagram doesn't tell us

[vandergus]

For example, I am currently looking for the function of the ratchet gear that is used in the Nikon F4 to control the aperture.

More precisely, the function of the coupled perforated disk that runs through a photointerrupter and generates pulses.

There's a similar component in Pentax cameras with automatic aperture control. I think it is observing and controlling how far the aperture stops down. Here's a more detailed explanation.

 

[Andreas Thaler]

Nevertheless, if we want to make progress with repairs, we must not give up.

Even if it is hard.

Many things seemed impossible to me at the beginning, but later I managed to do them. Some things didn't work.

But if no one tries, things will stay the way they are.

By the way, I chose my avatar here with consideration:

Austrian Archduke Carl, victor over Napoleon in the Battle of Aspern near Vienna, 1809.

The first to defeat Napoleon, something nobody had previously thought possible.

Archduke Karl monument

Photos, local info and history of this statue on Heldenplatz of the man who handed Napoleon his first major military defeat

www.visitingvienna.com

Archduke Charles, Duke of Teschen - Wikipedia

en.m.wikipedia.org

My ideal role model as a repairer

 

[Ian Leake]

My first job in about 1990 was as the field engineer for an oil company's facility in Libya. We still had 1970s era computers and tape decks built from discrete components. These things were an absolute nightmare to maintain. I broke as many things as I fixed. Thank goodness the world has moved on to integrated circuits, microcontrollers and C++.

 

[Andreas Thaler]

There's a similar component in Pentax cameras with automatic aperture control. I think it is observing and controlling how far the aperture stops down. Here's a more detailed explanation.

Yes, that's exactly how it should work in the F4, thank you!

 

[chuckroast]

I have been watching this thread with some interest. I am by training an electronics and communications engineer, though my career has diverged somewhat from that. At one point,
it was my job to have a technician decapsulate custom chips and then do a circuit analysis on what was exposed to confirm that none of my employer's patents were being violated by a competitor.

Without doing this, there is no reliable way to know what a magic chip may be doing. These days, it's even more difficult because of the subnanometer lithography being used for modern digital control, compute, and signal processing chips. You'd need very specialized and expensive equipment to be able to do that. That is to say, it's more-or-less impossible for most people (including people like me).

The F4 you're working on is a transition camera from traditional analog mostly controls, to higher degrees of automation for things like exposure and focus control. It's not surprising that Nikon would use custom chips to optimize this.

I guess what I'm saying is ... good luck figuring out all the magic boxes

 

[Andreas Thaler]

The F4 you're working on is a transition camera from traditional analog mostly controls, to higher degrees of automation for things like exposure and focus control. It's not surprising that Nikon would use custom chips to optimize this.

I guess what I'm saying is ... good luck figuring out all the magic boxes

I'm not completely crazy

Thank you and all for your comments.

For my part I'll concentrate on the block diagram of the F4 circuit.

The same applies to its mechanics.

The repair manual contains beautiful technical drawings. Enough to keep you guessing for a long time.

 

[BCM]

I'd also note that most circuit diagrams for consumer products are issued for the purpose of repair and not duplication.

 

[BrianShaw]

Context is also important. The circuit diagram in the OP is 1 page of a 176 page document, and 1 diagram in a multi-page explanation of the electrical circuit. It could be considered to be somewhat is disengenuous to take something like that out of context and expect anyone to decipher the functionality based on that diagram alone.

 

[Andreas Thaler]

Context is also important. The circuit diagram in the OP is 1 page of a 176 page document, and 1 diagram in a multi-page explanation of the electrical circuit. It could be considered to be somewhat disengenuous to take something like that out of context and expect anyone to decipher the functionality based on that diagram alone.

That's exactly what it's about, see the title of this thread: What a circuit diagram doesn't tell us.

 

[Laurent]

Around 1970.

In fact, on those Univac machines, it was also unusual for the user to replace discrete components. It's possible that the US Military had its own ways, but in the commercial world, this absolutely wasn't the way things were done back then. A service tech would come in, yank a module from the machine and put a new one in. These modules may then have been refurbished back at the manufacturer's facilities.

I worked on these in the 90s, while on the French Navy. We had complète diagrams at hand, the testing equipment to identify the faulty part (read oscilloscope and multimeter) and the tools and spare to replace said parts.

We also had sub-assemblies in limited inventory, but the discrete ICs and components where more universal.

During a repair, we could not only identify which IC failed, but also which part in it ( namely, which logic gate wasn't working as it should)

 

[koraks]

I worked on these in the 90s, while on the French Navy.

Yes, but that's quite an extraordinary environment. I've done research at a nuclear power plant where they also actively used and obviously maintained extremely well-documented control systems dating back to the 1960s. I'm sure they also replaced individual relays - yes, the control logic was implemented at least in part with relays, not even transistors. 1960s technology employing 1930s concepts. It's not the default, of course.

 

[BrianShaw]

.... and in those kinds of environments, whether really extraordinary or not, there is adequate education, training, and technical documentation to support the required tasks. Iv;e never experienced a situation where guesswork was the modus operandi except in reverse-engineering.

Before I retired I was a big proponent of "doing tomorrows mission with yesterday's technology" because that technology was well understood and proven. That solution only worked in situations where the mission was not one that rapidly evolved or iterated. The biggest downfall of that scheme is, and remains, supply chain for replacement parts. In somem situaitons the downfall is lack of full documenation when some was held by a vendor as proprietary and then they went out of business.

The 1990's were actually a long time ago, even though many of use remember them fondly like they were yesterday. The old way definitley isn't the default for modern systems, especially when they are commercial technology/equipment rather than purpose-built.

 

[Alan Edward Klein]

When did this end?

I worked for them from 1967-1970. During that time they already had some computers with IC's (Univac 1108) but the one I worked on (Univac 494) was all discrete components. The computers memories were bi-stable ferromagnetic cores, 30 bit octal words. Of course, storage was done with tape drives and large drums with surfaces made of magnetic nickel-cobalt.

 

[Alan Edward Klein]

Around 1970.

In fact, on those Univac machines, it was also unusual for the user to replace discrete components. It's possible that the US Military had its own ways, but in the commercial world, this absolutely wasn't the way things were done back then. A service tech would come in, yank a module from the machine and put a new one in. These modules may then have been refurbished back at the manufacturer's facilities.



Absolutely not, apart from perhaps very exceptional cases which will be hard to find. It's also likely that in 3d world countries component-level repairs are more usual given the low cost of labor in relation to parts cost. You'd see stuff like USB connectors being replaced etc.



It certainly is, and it only happens/happened in situations where there were very specific requirements. In the case of cameras, a driving force behind this was the need for extreme miniaturization (relative to what was common in the market at the time) and a high degree of complexity, resulting in highly integrated and specialized solutions. This was enabled by the manufacturing volumes and the high degree of vertical integration of the camera makers. This is one reason why companies like Nikon and Canon for some time actually kept up in semiconductor manufacturing (up until ca. 25 years ago).


As far as I'm concerned the main disagreement is that I try not to have expectations of other people that are unrealistic.

I worked in a Univac client's computer room which was thousands of square feet in size. Univac provided 24/7 on site coverage. So there was a lot of time to fix cards if the tech wanted to. Of course we had many spare circuit cards for swapping to get the machine back on line as quickly as possible. One of the more interesting repairs was when one of the many read/write heads on the Fastrand storage drums would "hit" the surface causing a retraction of all the 64? heads to protect the nickel cobalt surface. Then we'd have to disassemble the unit, and find the "hit" on the surface of the approx 20" diameter by 5 foot drum. We had to use a volcano pumice stone to smooth down the "hit" and then take those sectors out of storage as the process destroyed that portion of the surface.

 

[chuckroast]

I worked on these in the 90s, while on the French Navy. We had complète diagrams at hand, the testing equipment to identify the faulty part (read oscilloscope and multimeter) and the tools and spare to replace said parts.

We also had sub-assemblies in limited inventory, but the discrete ICs and components where more universal.

During a repair, we could not only identify which IC failed, but also which part in it ( namely, which logic gate wasn't working as it should)

I've done it both way. Swap-N-Fix or Diagnose-And-Fix

 

[chuckroast]

I worked for them from 1967-1970. During that time they already had some computers with IC's (Univac 1108) but the one I worked on (Univac 494) was all discrete components. The computers memories were bi-stable ferromagnetic cores, 30 bit octal words. Of course, storage was done with tape drives and large drums with surfaces made of magnetic nickel-cobalt.

Ahhhh the memories (so to speak). My first job out of college was doing field service on IBM 1130 3rd party disk controllers, drives, and printers. The company was happily surprised when I started fixing the cards we pulled out

 

[r_a_feldman]

How many people these days can fix a current-model car at the component level? Close to zero. All you can do is switch out a defective module with a new one. Even if you do something as simple as disconnect the battery, you need to reset the car’s computers to get them working again.

 

[BrianShaw]

How many people these days can fix a current-model car at the component level? Close to zero. All you can do is switch out a defective module with a new one. Even if you do something as simple as disconnect the battery, you need to reset the car’s computers to get them working again.

You are being too generous. The answer is zero.

 

[chuckroast]

You are being too generous. The answer is zero.

In fairness, the modules are not made to be serviced by anyone including factory techs. It's far more economical to build them in volume for swapout than for someone to laboriously work through component level analysis.

Moreover, with modern automated manufacturing and assembly, the reassembly of a fixed module would be nigh on impossible.

It's analogous to why we don't fix lightbulbs or rebuild oil filters on cars.

 

[reddesert]

I've done some building and repairing of electronic effect pedals for guitar and of music circuits (amplifiers, simple synthesizer modules, etc). At that level, it is possible to understand the signal flow using a schematic and diagnose problems using a voltmeter and sometimes an oscilloscope, if you have some basic knowledge of electronics and a sense of logical troubleshooting. This is with discrete components - resistors, capacitors, transistors, and simple common ICs like op-amps and occasional use of CMOS ICs, delay chips and so on.

However, it is a lot easier if you already know what the circuit is supposed to do. If you just look at the schematic without knowing whether it's a distortion pedal or an envelope follower, you have to figure that out and break the circuit into its components before diagnosing it.

People who work on such circuits without previous electronics training have a tendency to assume that the most complicated part of the circuit, such as an IC, is the likely source of problems. It's rarely the IC! Problems are far more commonly things like cold solder joints, bad switches, failed capacitors. Just as pandemic lockdown started, my cable modem failed. It's a device full of large custom surface mount ICs that one could never understand - but the part that failed was a 20 cent power supply filter capacitor, the one part I could replace in 5 minutes.

I took a basic electronics course in college, which was enough to teach me what an op-amp does for example. People without that training often think an op-amp is a device full of magic beans and guitar tonality. It's actually a precision high-gain amplifier that does one thing, drive its inputs to be equal. The internals of op-amps are an amazing work of decades of engineering discipline that you don't need to understand to use. The feedback circuit around the op-amp, a bunch of few-cent resistors, capacitors, and potentiometers, is what differentiates one audio circuit from another.

 

[Laurent]

Yes, but that's quite an extraordinary environment. I've done research at a nuclear power plant where they also actively used and obviously maintained extremely well-documented control systems dating back to the 1960s. I'm sure they also replaced individual relays - yes, the control logic was implemented at least in part with relays, not even transistors. 1960s technology employing 1930s concepts. It's not the default, of course.

I fully agree and measure the "privilege".

 

[Laurent]

I've done some building and repairing of electronic effect pedals for guitar and of music circuits (amplifiers, simple synthesizer modules, etc). At that level, it is possible to understand the signal flow using a schematic and diagnose problems using a voltmeter and sometimes an oscilloscope, if you have some basic knowledge of electronics and a sense of logical troubleshooting. This is with discrete components - resistors, capacitors, transistors, and simple common ICs like op-amps and occasional use of CMOS ICs, delay chips and so on.

However, it is a lot easier if you already know what the circuit is supposed to do. If you just look at the schematic without knowing whether it's a distortion pedal or an envelope follower, you have to figure that out and break the circuit into its components before diagnosing it.

People who work on such circuits without previous electronics training have a tendency to assume that the most complicated part of the circuit, such as an IC, is the likely source of problems. It's rarely the IC! Problems are far more commonly things like cold solder joints, bad switches, failed capacitors. Just as pandemic lockdown started, my cable modem failed. It's a device full of large custom surface mount ICs that one could never understand - but the part that failed was a 20 cent power supply filter capacitor, the one part I could replace in 5 minutes.

I took a basic electronics course in college, which was enough to teach me what an op-amp does for example. People without that training often think an op-amp is a device full of magic beans and guitar tonality. It's actually a precision high-gain amplifier that does one thing, drive its inputs to be equal. The internals of op-amps are an amazing work of decades of engineering discipline that you don't need to understand to use. The feedback circuit around the op-amp, a bunch of few-cent resistors, capacitors, and potentiometers, is what differentiates one audio circuit from another.

At the repair café where I'm spending some time, we have the same experience. "Failed" devices due to a faulty operator are fairly common (people forget about a safety mechanism and complain their wood machine does not start, this kind of things), and most of the rest is a failed fuse, power subsystem with a faulty cable, etc.

 

[chuckroast]

I've done some building and repairing of electronic effect pedals for guitar and of music circuits (amplifiers, simple synthesizer modules, etc). At that level, it is possible to understand the signal flow using a schematic and diagnose problems using a voltmeter and sometimes an oscilloscope, if you have some basic knowledge of electronics and a sense of logical troubleshooting. This is with discrete components - resistors, capacitors, transistors, and simple common ICs like op-amps and occasional use of CMOS ICs, delay chips and so on.

However, it is a lot easier if you already know what the circuit is supposed to do. If you just look at the schematic without knowing whether it's a distortion pedal or an envelope follower, you have to figure that out and break the circuit into its components before diagnosing it.

People who work on such circuits without previous electronics training have a tendency to assume that the most complicated part of the circuit, such as an IC, is the likely source of problems. It's rarely the IC! Problems are far more commonly things like cold solder joints, bad switches, failed capacitors. Just as pandemic lockdown started, my cable modem failed. It's a device full of large custom surface mount ICs that one could never understand - but the part that failed was a 20 cent power supply filter capacitor, the one part I could replace in 5 minutes.

I took a basic electronics course in college, which was enough to teach me what an op-amp does for example. People without that training often think an op-amp is a device full of magic beans and guitar tonality. It's actually a precision high-gain amplifier that does one thing, drive its inputs to be equal. The internals of op-amps are an amazing work of decades of engineering discipline that you don't need to understand to use. The feedback circuit around the op-amp, a bunch of few-cent resistors, capacitors, and potentiometers, is what differentiates one audio circuit from another.

The #1 thing I have to be faulty when diagnosing broken stuff is power related - power supplies, capacitors, even just bad fuses...

The #2 thing I have to be faulty when diagnosing broken stuff is mechanical - dirty contacts on switches and relays (Deoxit is your friend), faulty connectors and sockets...

The #3 thing I have to be faulty when diagnosing broken stuff is dirt and grime - dirty battery contacts, cooling fans and fins so loaded with dust they cannot function ..

I have #2 and #3 to frequently be issues when I am trying to rehab an old camera.

 

[Alan Edward Klein]

In fairness, the modules are not made to be serviced by anyone including factory techs. It's far more economical to build them in volume for swapout than for someone to laboriously work through component level analysis.

Moreover, with modern automated manufacturing and assembly, the reassembly of a fixed module would be nigh on impossible.

It's analogous to why we don't fix lightbulbs or rebuild oil filters on cars.

Reminds when I was a kid and we had vacuum tube TV's. When the thing went bad, you'd pull all the tubes and test them down at the nearby pharmacy that had a tube tester in the store along with replacement tubes to purchase.

 

[Alan Edward Klein]

I've done some building and repairing of electronic effect pedals for guitar and of music circuits (amplifiers, simple synthesizer modules, etc). At that level, it is possible to understand the signal flow using a schematic and diagnose problems using a voltmeter and sometimes an oscilloscope, if you have some basic knowledge of electronics and a sense of logical troubleshooting. This is with discrete components - resistors, capacitors, transistors, and simple common ICs like op-amps and occasional use of CMOS ICs, delay chips and so on.

However, it is a lot easier if you already know what the circuit is supposed to do. If you just look at the schematic without knowing whether it's a distortion pedal or an envelope follower, you have to figure that out and break the circuit into its components before diagnosing it.

People who work on such circuits without previous electronics training have a tendency to assume that the most complicated part of the circuit, such as an IC, is the likely source of problems. It's rarely the IC! Problems are far more commonly things like cold solder joints, bad switches, failed capacitors. Just as pandemic lockdown started, my cable modem failed. It's a device full of large custom surface mount ICs that one could never understand - but the part that failed was a 20 cent power supply filter capacitor, the one part I could replace in 5 minutes.

I took a basic electronics course in college, which was enough to teach me what an op-amp does for example. People without that training often think an op-amp is a device full of magic beans and guitar tonality. It's actually a precision high-gain amplifier that does one thing, drive its inputs to be equal. The internals of op-amps are an amazing work of decades of engineering discipline that you don't need to understand to use. The feedback circuit around the op-amp, a bunch of few-cent resistors, capacitors, and potentiometers, is what differentiates one audio circuit from another.

Back in the late 1950's or early 60s, building your own amp=preamp for a stereo system was the in thing to do. I built one from FisherKit of course with tubes. It worked for years and withstood shipping from NYC to Japan and back when I was in the USAF. I didn't understand electronics until later. But building it and seeing it actually work was a blast. The tuner and other peripherals I bought were built. Building once is enough!