Re: Transformer part ID please

[Adrian Graham]

On 24/01/2017 06:06, "Tony Duell"  wrote:

>> 
>> and it's 0.6ohm max.
> 
> Are you telling me that if you put 2 logic analyser inputs on 2 points on the
> same trace (which tests as continuous with an ohmmeter) that said 2 LA
> channels don't show the same thing? If so, the LA needs repairing!

...or binning given how much they cost. Thinking about it, with the 'simple
parallel' decoding option of the software I can clock the sampling using the
external 8085 clock on pin 37. If THAT gives inconsistent results then I can
only assume it's not up to the job.
 
> Given the problems I've had with even clean old sockets, I would have
> replaced those. Corroded sockets must be worse. And I can't believe
> they used turned-pin sockets in something like this.

Yes, it surprised me too. They do seem to have used decent components though
which is probably why it was so expensive back in 1984. I'll change the
sockets tonight.

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Tony Duell]

On Mon, Jan 23, 2017 at 11:38 PM, Adrian Graham
 wrote:
> On 23/01/2017 21:38, "Tony Duell"  wrote:

>> Oh... The trace isn't open-circuit, is it? Check it with the
>> multimeter. Given the
>> corrosion damage to the tape drive I could well believe PCB problems of a
>> similar nature.
>
> I've measured resistance from all points to all points, there's 9:
>
> 74LS373 (output, new socket checked for solder bridges, new chip)
> 4x 2764 EPROM (new chips, no new sockets yet)
> MC3242A (new chip, no new socket yet, this one had bad verdigris)
> 74LS139
> 74LS21
> 74LS365
>
> and it's 0.6ohm max.

Are you telling me that if you put 2 logic analyser inputs on 2 points on the
same trace (which tests as continuous with an ohmmeter) that said 2 LA
channels don't show the same thing? If so, the LA needs repairing!

Given the problems I've had with even clean old sockets, I would have
replaced those. Corroded sockets must be worse. And I can't believe
they used turned-pin sockets in something like this.

-tony

Re: Transformer part ID please

[Adrian Graham]

On 23/01/2017 21:38, "Tony Duell"  wrote:

> I normally trace connections using my DMM, but of course you have to know
> what will fool it (i.e. low resistance components). I find I have to desolder
> relays, switches, inductors, transformers, low value resistors, etc before
> starting to trace the circuit.

There's been a lot of that here too.

>> should go high at the same time as long as it's part of the same trace, so
>> the LS373 to MC3242A to pin 3 on all four ROMs should pulse together and
>> they don't. If there's a short they should all get it.
> 
> Oh... The trace isn't open-circuit, is it? Check it with the
> multimeter. Given the
> corrosion damage to the tape drive I could well believe PCB problems of a
> similar nature.

I've measured resistance from all points to all points, there's 9:

74LS373 (output, new socket checked for solder bridges, new chip)
4x 2764 EPROM (new chips, no new sockets yet)
MC3242A (new chip, no new socket yet, this one had bad verdigris)
74LS139
74LS21
74LS365

and it's 0.6ohm max.

I picked A7 because it and its multiplexed D7 are special cases and go to
quite a few locations. D0-D6's input to the teletext chip from the CPU is
via a 74LS240, D7's input is based on ALE/A6/A7 and a few other things
involving clocks that aren't yet present (lack of SOD and IO/M) so I'm
fairly sure the teletext chip isn't getting a full byte.

Both my analysers are cheap chinese clones, I suspect I need something
beefier.

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Tony Duell]

On Mon, Jan 23, 2017 at 9:23 PM, Adrian Graham
 wrote:
> On 23/01/2017 19:07, "Tony Duell"  wrote:
>
>>> I didn't think of a 555 in that scenario but that makes sense. I've marked
>>> it as such for now, cheers!
>>
>> Is there another capacitor from the 2/6 junction to chassis? Any other
>> components
>> around? It might be one of the special SMPSU ICs, but it doesn't
>> instantly ring a
>> bell with me.
>
> Yes, just found it after some leg cleaning. This board is a big fan of
> running small traces under other components. Looks like it's a 68pF ceramic
> type.

Sounds reasonable...

I normally trace connections using my DMM, but of course you have to know
what will fool it (i.e. low resistance components). I find I have to desolder
relays, switches, inductors, transformers, low value resistors, etc before
starting to trace the circuit.

>
>>> Tonight it's 'address pins and why I get inconsistent results on a logic
>>> analyser while looking at everything that touches A7' :)
>>
>> Short from A7 to another pin? Failed address latch?
>
> I've been looking for shorts but I'd have thought that anything that's A7
> should go high at the same time as long as it's part of the same trace, so
> the LS373 to MC3242A to pin 3 on all four ROMs should pulse together and
> they don't. If there's a short they should all get it.

Oh... The trace isn't open-circuit, is it? Check it with the
multimeter. Given the
corrosion damage to the tape drive I could well believe PCB problems of a
similar nature.

-tony

Re: Transformer part ID please

[Adrian Graham]

On 23/01/2017 19:07, "Tony Duell"  wrote:

>> I didn't think of a 555 in that scenario but that makes sense. I've marked
>> it as such for now, cheers!
> 
> Is there another capacitor from the 2/6 junction to chassis? Any other
> components
> around? It might be one of the special SMPSU ICs, but it doesn't
> instantly ring a
> bell with me.

Yes, just found it after some leg cleaning. This board is a big fan of
running small traces under other components. Looks like it's a 68pF ceramic
type.

>> Tonight it's 'address pins and why I get inconsistent results on a logic
>> analyser while looking at everything that touches A7' :)
> 
> Short from A7 to another pin? Failed address latch?

I've been looking for shorts but I'd have thought that anything that's A7
should go high at the same time as long as it's part of the same trace, so
the LS373 to MC3242A to pin 3 on all four ROMs should pulse together and
they don't. If there's a short they should all get it.

Since this is an 8085A powered machine it runs internally at 3MHz so I
should be able to sample it at over 12MS/s. Another collector friend is
loaning me an analyser that can be externally clocked but it's still USB
connected like both of mine are. I'm pondering an HP/Agilent 1663A that's on
ebay for 'offers'...

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Tony Duell]

On Mon, Jan 23, 2017 at 6:48 PM, Adrian Graham
 wrote:
> On 22/01/2017 20:10, "Tony Duell"  wrote:
>
>>> That looks like an opamp to me, with bias pins connected to GND?
>>
>> Does it? It looks like a 555 timer to me at first glance. I would have
>> expected another
>> capacitor from pins 2/6 to ground though.
>>
>> 1 Ground
>> 2 Trigger (linked to threshold, and to output via a timing resistor)
>> 3 Output (driving the MOSFET)
>> 4 Reset/ (pulled high via a resistor)
>> 5 Control Voltage (decoupled to ground)
>> 6 Threshold (see 2)
>> 7 Discharge (not used here)
>> 8 Vcc
>
> I didn't think of a 555 in that scenario but that makes sense. I've marked
> it as such for now, cheers!

Is there another capacitor from the 2/6 junction to chassis? Any other
components
around? It might be one of the special SMPSU ICs, but it doesn't
instantly ring a
bell with me.


>
> Tonight it's 'address pins and why I get inconsistent results on a logic
> analyser while looking at everything that touches A7' :)

Short from A7 to another pin? Failed address latch?

-tony

Re: Transformer part ID please

[Adrian Graham]

On 22/01/2017 20:10, "Tony Duell"  wrote:

>> That looks like an opamp to me, with bias pins connected to GND?
> 
> Does it? It looks like a 555 timer to me at first glance. I would have
> expected another
> capacitor from pins 2/6 to ground though.
> 
> 1 Ground
> 2 Trigger (linked to threshold, and to output via a timing resistor)
> 3 Output (driving the MOSFET)
> 4 Reset/ (pulled high via a resistor)
> 5 Control Voltage (decoupled to ground)
> 6 Threshold (see 2)
> 7 Discharge (not used here)
> 8 Vcc

I didn't think of a 555 in that scenario but that makes sense. I've marked
it as such for now, cheers!

Tonight it's 'address pins and why I get inconsistent results on a logic
analyser while looking at everything that touches A7' :)

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Tony Duell]

> Result! Whilst photographing the board to trace out that circuit (it's
> quicker for me than drawing it) I removed all the big capacitors. While they
> were out of circuit I tested them and discovered that not only was one dead
> in the -12V rail but also the through hole plating for another on the +12V
> rail had failed to a trace on the back of the board so neither the unknown
> 8-pin DIP or the LF351N were getting +12V.

Excellent!

>
> Fixed that and replaced the failed cap and now have -24V where it should be.
>
> If it helps with identification the unknown chip appears to be:
>
> 1 - GND
> 2 - 56k resistor to pin 3
> 3 - Gate of IRF523
> 4 - +12V via 10k resistor
> 5 - GND via decoupling cap
> 6 - connected to 2
> 7 - NC
> 8 - +12V
>
> That looks like an opamp to me, with bias pins connected to GND?

Does it? It looks like a 555 timer to me at first glance. I would have
expected another
capacitor from pins 2/6 to ground though.

1 Ground
2 Trigger (linked to threshold, and to output via a timing resistor)
3 Output (driving the MOSFET)
4 Reset/ (pulled high via a resistor)
5 Control Voltage (decoupled to ground)
6 Threshold (see 2)
7 Discharge (not used here)
8 Vcc


-tony

Re: Transformer part ID please

[Adrian Graham]

On 22/01/2017 07:31, "Tony Duell"  wrote:

>> The -24V regulator can be seen just above the 4 blue caps, The regulator
>> that can be seen under the black wire is a 7812. The LF351N has -12V present
>> at its bias outputs, the opamp to its right has no markings helpfully.
> 
> What about that IRF523? That is a power MOSFET transistor and might well
> be the chopper in the switching regulator circuit.

Result! Whilst photographing the board to trace out that circuit (it's
quicker for me than drawing it) I removed all the big capacitors. While they
were out of circuit I tested them and discovered that not only was one dead
in the -12V rail but also the through hole plating for another on the +12V
rail had failed to a trace on the back of the board so neither the unknown
8-pin DIP or the LF351N were getting +12V.

Fixed that and replaced the failed cap and now have -24V where it should be.

If it helps with identification the unknown chip appears to be:

1 - GND
2 - 56k resistor to pin 3
3 - Gate of IRF523
4 - +12V via 10k resistor
5 - GND via decoupling cap
6 - connected to 2
7 - NC
8 - +12V

That looks like an opamp to me, with bias pins connected to GND?

One more failed component off the list anyway.

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Adrian Graham]

On 22/01/2017 07:31, "Tony Duell"  wrote:

>>> It looks to be part of a switching regulator circuit. Possibly to generate
>>> other voltages from the 5V line. Can you post a picture of more of the PCB
>>> around it.
>> 
>> Its second output goes via an IN4148 to the -24V regulator:
> 
> What do you mean by 'second output'? How many pins are used on this
> 'transformer'? If only 2 it is just a simple inductor, again possibly part of
> a swtiching regulator.

There's 5 pins (5th for keying I guess) with 2 separate windings and I've
just realised the 2nd winding (destination -24V) isn't fed from anywhere.
It's only a 2 sided board and there are no top traces under the part.

There's also some incomplete traces in that circuit so I wonder if there
were planned enhancements for a Mark II that never happened. I'm marking up
the back of the board for my own understanding as I type.

>> The -24V regulator can be seen just above the 4 blue caps, The regulator
>> that can be seen under the black wire is a 7812. The LF351N has -12V present
>> at its bias outputs, the opamp to its right has no markings helpfully.
> 
> What about that IRF523? That is a power MOSFET transistor and might well
> be the chopper in the switching regulator circuit.

It tests OK in my MOSFET tester.
 
> Why do you assume the other 8 pin IC is an op-amp?

Initially because of the way it's wired but I'm checking it now. Once I've
marked up where the traces go the picture should be clearer.

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Tony Duell]

On Sat, Jan 21, 2017 at 11:33 PM, Adrian Graham
 wrote:
> On 21/01/2017 21:59, "Tony Duell"  wrote:
>
>> On Sat, Jan 21, 2017 at 8:43 PM, Adrian Graham
>>  wrote:
>>> Hi folks,
>>>
>>> Does anyone know what this is? It's obviously a transformer coil(s) but
>>> image search and googling those markings produces nothing:
>>>
>>> http://www.binarydinosaurs.co.uk/la4543.jpg
>>
>> I think the LA number is the number for the ferrite 'pot core' not the
>> complete transformer.
>>
>> It looks to be part of a switching regulator circuit. Possibly to generate
>> other voltages from the 5V line. Can you post a picture of more of the PCB
>> around it.
>
> Its second output goes via an IN4148 to the -24V regulator:

What do you mean by 'second output'? How many pins are used on this
'transformer'? If only 2 it is just a simple inductor, again possibly part of
a swtiching regulator.

>
> http://www.binarydinosaurs.co.uk/STCExecutelPowerRegulation.jpg
>
> The top black wire is from the PSU and goes left-right (in pairs) GND, +5,
> +12, -12. I'd do an underside pic too but the board's covered in analyser
> wires at the mo.
>
> The -24V regulator can be seen just above the 4 blue caps, The regulator
> that can be seen under the black wire is a 7812. The LF351N has -12V present
> at its bias outputs, the opamp to its right has no markings helpfully.

What about that IRF523? That is a power MOSFET transistor and might well
be the chopper in the switching regulator circuit.

Why do you assume the other 8 pin IC is an op-amp?

-tony

Re: Transformer part ID please

[Adrian Graham]

On 21/01/2017 21:59, "Tony Duell"  wrote:

> On Sat, Jan 21, 2017 at 8:43 PM, Adrian Graham
>  wrote:
>> Hi folks,
>> 
>> Does anyone know what this is? It's obviously a transformer coil(s) but
>> image search and googling those markings produces nothing:
>> 
>> http://www.binarydinosaurs.co.uk/la4543.jpg
> 
> I think the LA number is the number for the ferrite 'pot core' not the
> complete transformer.
> 
> It looks to be part of a switching regulator circuit. Possibly to generate
> other voltages from the 5V line. Can you post a picture of more of the PCB
> around it.

Its second output goes via an IN4148 to the -24V regulator:

http://www.binarydinosaurs.co.uk/STCExecutelPowerRegulation.jpg

The top black wire is from the PSU and goes left-right (in pairs) GND, +5,
+12, -12. I'd do an underside pic too but the board's covered in analyser
wires at the mo.

The -24V regulator can be seen just above the 4 blue caps, The regulator
that can be seen under the black wire is a 7812. The LF351N has -12V present
at its bias outputs, the opamp to its right has no markings helpfully.

Cheers,

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Tony Duell]

On Sat, Jan 21, 2017 at 8:43 PM, Adrian Graham
 wrote:
> Hi folks,
>
> Does anyone know what this is? It's obviously a transformer coil(s) but
> image search and googling those markings produces nothing:
>
> http://www.binarydinosaurs.co.uk/la4543.jpg

I think the LA number is the number for the ferrite 'pot core' not the
complete transformer.

It looks to be part of a switching regulator circuit. Possibly to generate
other voltages from the 5V line. Can you post a picture of more of the PCB
around it.

-tony

Re: Transformer part ID please

[Adrian Graham]

On 21/01/2017 21:08, "Chuck Guzis"  wrote:

> On 01/21/2017 12:43 PM, Adrian Graham wrote:
>> Hi folks,
>> 
>> Does anyone know what this is? It's obviously a transformer coil(s)
>> but image search and googling those markings produces nothing:
>> 
>> http://www.binarydinosaurs.co.uk/la4543.jpg
>> 
>> It's taking and giving +5V but I'm sure it should also be producing
>> another voltage...
> 
> The marking identifies it as a Mullard ferrite core (LA4543) but says
> nothing about how it's wound.

I kinda hoped it was a stock part given that everything else in this machine
is right off the shelf or at least appears to be...

Cheers!

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: Transformer part ID please

[Jon Elson]

On 01/21/2017 02:43 PM, Adrian Graham wrote:

Hi folks,

Does anyone know what this is? It's obviously a transformer coil(s) but
image search and googling those markings produces nothing:

http://www.binarydinosaurs.co.uk/la4543.jpg

It's taking and giving +5V but I'm sure it should also be producing another
voltage...

Cheers!

That is a ferrite pot core.  They are very often custom 
devices made for one specific purpose.  Pulse Engineering 
and a few other outfits have stock parts like that, but if 
you can't match the part number, it is likely to either be a 
house-numbered part or a custom.


Jon

Re: Transformer part ID please

[Chuck Guzis]

On 01/21/2017 12:43 PM, Adrian Graham wrote:
> Hi folks,
> 
> Does anyone know what this is? It's obviously a transformer coil(s)
> but image search and googling those markings produces nothing:
> 
> http://www.binarydinosaurs.co.uk/la4543.jpg
> 
> It's taking and giving +5V but I'm sure it should also be producing
> another voltage...

The marking identifies it as a Mullard ferrite core (LA4543) but says
nothing about how it's wound.

--Chuck

Transformer part ID please

[Adrian Graham]

Hi folks,

Does anyone know what this is? It's obviously a transformer coil(s) but
image search and googling those markings produces nothing:

http://www.binarydinosaurs.co.uk/la4543.jpg

It's taking and giving +5V but I'm sure it should also be producing another
voltage...

Cheers!

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?