Commodore 8026 printer / typewriter

[TonyDuell]

I think a base-collector short in the associated transistor would give the odd readings. 0.6V or so when the diode is forward biased. And the b-e juncton + the clamp zener ('backwards') gving 2 diode drops or 1.2V when the diode in question is reverse-biased

[TonyDuell]

Originally Posted by SiriusHardware

Question for Tony: The 'Home' sensor circuit is slightly more complicated than I anticipated - Q3 / Q4 looks like a transistor multivibrator circuit influenced by the 'home' sensor, so it looks as though, rather than generating a static 'high' or 'low' signal to feed to the INT input, it will generate a pulse train. Can you walk us through the operation of that part?

Yes, that's what it looks like to me too. Seems a bit odd, but that's what the circuit was. And the Olympia ES100 typewriter is much the same circuit.

Please remember I don't own such a printer and have never seen one working, so I have no more information than you do

[SiriusHardware]

You have vastly more experience with, and understanding of, 'hybrid' electronic / mechanical gear than I do. My experience tends more towards the purely electronic.

I'm trying to decide whether the input from the 'home' sensor actually disables / enables the multivibrator or only alters its frequency / pulse width. The 'missing' 10K collector resistor for Q3 is presumably present in the form of an internal pull-up on the INT pin of U9.

Maybe the INT input triggers on a falling edge and the multivibrator is there to make sure it definitely sees at least one falling edge when the 'home' position is reached. I also considered that it might be a one-shot, (monostable) to generate a reliable single pulse of known width when the home sensor 'detects', but with the two 4n7 caps rather than just one, it looks more like an astable.

None of this matters much as long as the INT pin sees one thing when the carriage is 'homed', and another when it is not.

[TonyDuell]

That's why I asked Colin 'what do you see on the INT/ pin'. Provided there are 2 distinct signals, one when the carriage is home, the other when it's not, it's likely the sensor is working correctly.

I am happy to carry on contributing to this thread and to give my thoughts. All I was sayig was that they are just my thoughts/deductions from the circuit diagrams, etc. Not actual measurements from a working unit. And I most definitely make mistakes!

[SiriusHardware]

I think a base-collector short in the associated transistor would give the odd readings.

Sounds like a promising lead although I am surprised that didn't show up somehow in the readings when Colin checked the base-collector resistances of the four transistors earlier - although what you're suggesting is a base-emitter short on the second transistor of the internal transistor pair?

Colin could, I suppose, remove it and try it on his new fangled tester box: Then try the same thing with any one of the other three to see whether it reads / tests the same or differently.

[ScottishColin]

OK - due to either

a) my ineptitude or
b) the diode being old and knackered

(I'm going with a and sticking to it) the diode has come apart.

Please can someone let me know what diode I need to replace it with?

Thanks.

Colin.

Originally Posted by SiriusHardware

Checks for today:

If possible, re-test that diode out-of-circuit or at least desoldered / disconnected at one end. If it then measures the same as the others (0.64V / 'OV'), reinstate it and we will consider other possible reasons for that different measurement when in-circuit. If it still measures differently out of circuit, stop and let us know.

Check what happens on the INT pin (pin 38) of U9, first with the carriage NOT in the 'homed' position and then with the carriage in the 'homed' position. If the INT pin shows one state when homed and another when not homed, that is what we would expect. If you don't see that behaviour, report back.

With the carriage centred (not homed), for each of U9 pins 33, 32, 31, 30, scope what happens on those pins immediately following switch-on. If the drive signals (digital waveforms) are present there try following them through the relevant gates of U1 to the bases of Q7, Q8, Q9, Q10.

[SiriusHardware]

It would be unusual for a plastic diode to just fall apart like that so maybe it really was faulty, although see Tony's insight (#261) into a possible problem with the associated transistor, and not the diode after all.

Are there any numbers readable on the other diodes? They will all be the same. The good news is that these will just be ordinary silicon rectifier diodes so anything with a similar voltage and current rating should do, but ideally you want to replace with like for like.

[SiriusHardware]

I found the link in post #109 to a place where the schematics for the underlying Typewriter are held: Unfortunately I can't unzip the schematic here at work, as we have nothing on our locked-down IT which will unpack 7zipped files.

Can someone have a look and see if the diodes in question (the ones connected to Q7, Q8, Q9, Q10 collector) are identified by type / part number on that diagram?

[SiriusHardware]

although what you're suggesting is a base-emitter short on the second transistor of the internal transistor pair?

Sorry, this was a finger fumble - you said base-collector short, of course.

[TonyDuell]

Originally Posted by SiriusHardware

I found the link in post #109 to a place where the schematics for the underlying Typewriter are held: Unfortunately I can't unzip the schematic here at work, as we have nothing on our locked-down IT which will unpack 7zipped files.

Can someone have a look and see if the diodes in question (the ones connected to Q7, Q8, Q9, Q10 collector) are identified by type / part number on that diagram?

It's not very readable, but it almost looks like 1N4148 which surprises me.

Colin, can you post a photo of the diode in question and a still-intact one on the board, please.

[ScottishColin]

Here's a photo as requested.

Colin.

[ScottishColin]

I've zoomed in and I'm reading 1M6168?

Colin.

Originally Posted by TonyDuell

It's not very readable, but it almost looks like 1N4148 which surprises me.

Colin, can you post a photo of the diode in question and a still-intact one on the board, please.

[TonyDuell]

Originally Posted by ScottishColin

Here's a photo as requested.

Colin.

Thanks...

It looks like yellow (wide), brown, yellow, grey stripes. Which would be '4148'.

The 1N4148 is a very common diode but I am surprised it has a high enough current rating for this applicatiion.

[TonyDuell]

Colin, could you now do the diode test on the non-existant diode (!). What I mean is to use the diode test range on your multimeter and put the probes on the 2 pads the diode came from, both ways round. That's with the unit turned off and J1 disconnected, of course.

I have my suspicions as to what the readings will be but don't want to influence your measurements...

[Mark1960]

I remember using 1N4448 where the 1N4148 couldn’t handle the current, same size but marked yellow yellow yellow grey. It might be worth using as a replacement if you don’t already have 1N4148 available.

[SiriusHardware]

The colour bands are the old way of marking those diodes, they are still in common use today, modern replacements more usually have the numbers 4148 printed on them but they are the same. Like Tony I am surprised at such a tiddly little diode being used for this purpose.

With the diode now well and truly out of the way the measurement he suggests should show whether his suspicion about the Darlington is correct.

[ScottishColin]

I cannot get a reading in either direction (ie Overload).

I have tested continuity onwards for both pads and that's OK so I believe that the readings are correct.

Is that what you suspected?

Colin.

Originally Posted by TonyDuell

Colin, could you now do the diode test on the non-existant diode (!). What I mean is to use the diode test range on your multimeter and put the probes on the 2 pads the diode came from, both ways round. That's with the unit turned off and J1 disconnected, of course.

I have my suspicions as to what the readings will be but don't want to influence your measurements...

[SiriusHardware]

Is that what you suspected?

Probably not, I think Tony was hoping to see the 1.2V reverse voltage drop even with the diode out.

At the moment we are probably back to suspecting the diode was in fact faulty, although that 1.2V reverse voltage is suspiciously close to what we'd expect from a double-diode.

Order some 1N4148 diodes anyway, as you are obviously going to have to replace the damaged one. When the new one is in you can try that measurement again with the new diode in place, see what we get then.

[ortek_service]

Originally Posted by Mark1960

I remember using 1N4448 where the 1N4148 couldn’t handle the current, same size but marked yellow yellow yellow grey. It might be worth using as a replacement if you don’t already have 1N4148 available.

Well Vishay at least have identical max ratings for their 1N4148 & 1N4448's:

https://www.vishay.com/docs/81857/1n4148.pdf
https://www.vishay.com/docs/81858/1n4448.pdf

- Makes you wonder why they make both of these, unless it is just to be a second-source for other manufacturers
And they do also have separate datasheets (although on the 1N4148 type they say it is also electrically-equivalent to the 1N914)
However, they do only give Vf at 10mA for the 1N4148, whereas they give Vf at 5mA + 100mA for the 1N4448.

For both of these their max ratings include:
Peak forward surge current tp = 1 ?s IFSM 2 A
Repetitive peak forward current IFRM 500 mA
Forward continuous current IF 300 mA
Average forward current VR = 0 IF(AV) 150 mA

- I always thought 1N4148's were only rated at 100mA continuous / average , so maybe Vishay make higher-spec ones?

So it seems that if there are high peak-currents, then the max. average current rating is strangely half that of the max. continuous current rating.

Not sure if 1N4448's are as cheap as 1N4148's (used to be <=4p each, and I've often bought packs of 100 for a £1 or so)
But as diodes are normally rather cheaper than the postage, then it may be worth obtaining some spares for all the different types used in this equipment in case any others get damages / common ones are always useful for building various projects.

[TonyDuell]

Originally Posted by SiriusHardware

Probably not, I think Tony was hoping to see the 1.2V reverse voltage drop even with the diode out.

At the moment we are probably back to suspecting the diode was in fact faulty, although that 1.2V reverse voltage is suspiciously close to what we'd expect from a double-diode.

Yes, I was expecting 1.2V 'backwards'. So the original measurements would have been the 0.6V from a good diode and 1.2V from other things on the board.

It just proves you should not assume things. Always make the measurement.

I'd fit a new diode and then re-do the tests.