UK fixer for desperate hobbyist?

[Terry Pinnell]

This is a longshot, but I'm sort of clutching at straws...

A year or so ago, I bought a second-hand function generator on ebay,
and it's just developed a fatal flaw. It's a Philips (now Fluke)
PM5134, an impressive instrument originally costing a couple of
thousand I think, and which I got for a couple of hundred. The DC
Offset is broken. Not only does it now get nowhere near zero, but the
error is so large that it distorts all waveforms above a small
amplitude. I was fairly confident that Fluke could do *something*
with it, at a price. I spent a couple of hours packing it and sent it
for via ParcelForce to Norwich (Ł17, it totalled about 8kg). So I'm
mortified to get this reply:

"With reference to the above instrument which you have sent in for
investigation and calibration, the unit has been thoroughly examined
by our technicians and the unit has found to be beyond repair."

Frankly, I suspect Fluke took one look at its age, checked that it was
discontinued, and abandoned it. I'll try getting more info by phone
tomorrow, but I'm not hopeful. Anyway, I'm awaiting its return now.

I had already had the covers off, pored over the circuit diagram,
tentatively tried a few presets and measured a few possibly relevant
voltages. (I bought the comprehensive user manual with circuit and
calibration instructions for a song from Fluke ages ago.) But I'm
really out of my depth. I *think* it's probably a 'simple' power
supply failure, but I'm nervous about replacing components.

So, I'm wondering if anyone can recommend a potential 'fixer' please?
Someone, amateur or professional, armed with the manual & schematics,
who would have a go at it. Apart from delivery, I'd pay say Ł100 to
get it back in good working order, or say Ł30 for a genuinely good
try but no seegar, i.e. 'Fluke were right, we can't fix it either'.

It seems such a shame to just dump it ;-(

--
Terry Pinnell
Hobbyist, West Sussex, UK

 

[Nigel Cook]

"Terry Pinnell" <terrypinDELETE@dial.pipexTHIS.com> wrote in message
news:144bhvo39a4b8pruhs6cii9ue98tmt2gvj@4ax.com...

This is a longshot, but I'm sort of clutching at straws...

A year or so ago, I bought a second-hand function generator on ebay,
and it's just developed a fatal flaw. It's a Philips (now Fluke)
PM5134, an impressive instrument originally costing a couple of
thousand I think, and which I got for a couple of hundred. The DC
Offset is broken. Not only does it now get nowhere near zero, but the
error is so large that it distorts all waveforms above a small
amplitude. I was fairly confident that Fluke could do *something*
with it, at a price. I spent a couple of hours packing it and sent it
for via ParcelForce to Norwich (Ł17, it totalled about 8kg). So I'm
mortified to get this reply:

"With reference to the above instrument which you have sent in for
investigation and calibration, the unit has been thoroughly examined
by our technicians and the unit has found to be beyond repair."

Frankly, I suspect Fluke took one look at its age, checked that it was
discontinued, and abandoned it. I'll try getting more info by phone
tomorrow, but I'm not hopeful. Anyway, I'm awaiting its return now.

I had already had the covers off, pored over the circuit diagram,
tentatively tried a few presets and measured a few possibly relevant
voltages. (I bought the comprehensive user manual with circuit and
calibration instructions for a song from Fluke ages ago.) But I'm
really out of my depth. I *think* it's probably a 'simple' power
supply failure, but I'm nervous about replacing components.

So, I'm wondering if anyone can recommend a potential 'fixer' please?
Someone, amateur or professional, armed with the manual & schematics,
who would have a go at it. Apart from delivery, I'd pay say Ł100 to
get it back in good working order, or say Ł30 for a genuinely good
try but no seegar, i.e. 'Fluke were right, we can't fix it either'.

It seems such a shame to just dump it ;-(

--
Terry Pinnell
Hobbyist, West Sussex, UK

If you can personally deliver it to me across the border -
not post or courier I can look at it.
25 years experience of repair of all sorts of kit
including scopes and assorted test equipment.

e-mail (removing .....) diverse2@tcp.co.u.....k
electronic hints and repair briefs , schematics/manuals list on
http://homepages.tcp.co.uk/~diverse

Nigel,Diverse Devices,Southampton,England

 

[Paul Burridge]

On Thu, 17 Jul 2003 01:09:25 +0100, "John Jardine"
<john@jjdesigns.fsnet.co.uk> wrote:

Terry Pinnell <terrypinDELETE@dial.pipexTHIS.com> wrote in message
news:144bhvo39a4b8pruhs6cii9ue98tmt2gvj@4ax.com...
This is a longshot, but I'm sort of clutching at straws...
[clip]
--
Terry Pinnell
Hobbyist, West Sussex, UK

At a guess ...
The fault's going to be in the output amplifier. Essentially it'll be a DC
coupled 'Hi-fi style' power amp (gain of about X10) able to drive a 50ohm
load. One (or more) of the ouput transistors has blown and the amp is trying
to work on only half of the input signal. It should be fixable unless
they've used some special I.C.
Maybe post the circuit with a couple of DCV measurements drawn on?. I'd
shudder at the thought of junking it.

Yes, it's always a shame. I've a retired electronic serviceman and ham
friend nearby who'd probably be happy to take this on little job if
Terry's stuck.

 

[Terry Pinnell]

"Nigel Cook" <diverse2@tcp.co.antispam.uk> wrote:

If you can personally deliver it to me across the border -
not post or courier I can look at it.
25 years experience of repair of all sorts of kit
including scopes and assorted test equipment.

e-mail (removing .....) diverse2@tcp.co.u.....k
electronic hints and repair briefs , schematics/manuals list on
http://homepages.tcp.co.uk/~diverse

Nigel,Diverse Devices,Southampton,England

Many thanks, Nigel, much appreciate the offer. I posted simultaneously
on UK's CIX Forum and had similar offer late last night from someone
just a few miles away in Crowborough, which I promptly accepted. Hope
to get the unit back from Fluke before my 2 weeks hols on Friday and
if so will deliver it to him with manual.

Regards,

--
Terry, East Grinstead, West Sussex, UK

 

[Terry Pinnell]

Paul Burridge <redscar@waitrose.notformail.com> wrote:

On Thu, 17 Jul 2003 01:09:25 +0100, "John Jardine"
john@jjdesigns.fsnet.co.uk> wrote:


Terry Pinnell <terrypinDELETE@dial.pipexTHIS.com> wrote in message
news:144bhvo39a4b8pruhs6cii9ue98tmt2gvj@4ax.com...
This is a longshot, but I'm sort of clutching at straws...
[clip]
--
Terry Pinnell
Hobbyist, West Sussex, UK

At a guess ...
The fault's going to be in the output amplifier. Essentially it'll be a DC
coupled 'Hi-fi style' power amp (gain of about X10) able to drive a 50ohm
load. One (or more) of the ouput transistors has blown and the amp is trying
to work on only half of the input signal. It should be fixable unless
they've used some special I.C.
Maybe post the circuit with a couple of DCV measurements drawn on?. I'd
shudder at the thought of junking it.

Yes, it's always a shame. I've a retired electronic serviceman and ham
friend nearby who'd probably be happy to take this on little job if
Terry's stuck.

Thanks, Paul. Could have been the same chap, except that my contact is
apparently still working...

--
Terry Pinnell
Hobbyist, West Sussex, UK

 

[John Crighton]

On Fri, 18 Jul 2003 10:17:33 +0100, Terry Pinnell
<terrypinDELETE@dial.pipexTHIS.com> wrote:

Thanks for the confidence-booster, John!

I suppose a bunch of factors prompted my cry for help:

- When it comes to fixing my *own* kit, be it work bench stuff like
this or my PCs, for reasons I've never understood any IQ, confidence
and skills I have take a dive, and my nervousness increases
proportionally!

- I'd had the covers off a few times and spent many hours, in vain.
Some months ago I'd managed to re-calibrate a couple of things, but
this gross flaw was clearly not going to be a push-over.

- Access is a problem. Its original cost, (many years ago, shortly
after the discovery of electricity), was over 3,000 UKP, and I'm
scared of making things terminally worse.

But I do like the idea of "Bob Parker's little electrolytic capacitor
checker". Actually, I suspect a dual power supply output transistor
has gone, as John Jardine suggested, but I'd like to be able to give
the key caps a checkout first as you recommend.

Even a few minutes research plainly shows that (quite apart from the
immediate issue of my function generator), I need to learn more about
ESR. Not sure how I've managed to escape its implications for so many
years. For example, this typical quote from
http://www.flippers.com/esrkttxt.html was an eye-opener to me:
"In fact the service life of electrolytic capacitors is approximately
halved for every 10 degrees C increase in temperature, and I was very
surprised to learn that many electrolytics are designed for a reliable
operating life of only a few thousand hours at their maximum rated
temperature and ripple current. (Remember that a single year is 8,766
hours!)"

I made an In Situ Transistor Tester many years ago, which has proved
invaluable.
(http://www.terrypin.dial.pipex.com/Images/InSituTransistorTester.gif)
Looks like I may need to make, or even buy an equivalent ESR tester.

--
Terry Pinnell
Hobbyist, West Sussex, UK

Hello Terry,
lets start with the easy stuff.
Check all the various power supply rails with your
voltmeter and CRO. If you can't reach a test point
then solder some wire extentions.

You mentioned the output stage as a possible
faulty area, well lets start snipping components,
testing them and solder blobing the snipped leads
together after they have been tested. Sounds
rough but that's how your telly gets fixed. We are
not working for NASA, we just want this unit to
go again. Pull out all the transistors associated
with the output stage and check them. If you can
use that "in circuit tester" of yours, great.

Process of elimination. You say to yourself, "I know
this is good and that is good because I checked
them. OK what is left?"
Hoe into it Terry, I know you can do it.
After you find the fault, all the snipped
and solder blobed components can be
replaced later if you wish.

Regards,
John Crighton
Sydney

 

[Roger Johansson]

Terry Pinnell <terrypinDELETE@dial.pipexTHIS.com> wrote:

A year or so ago, I bought a second-hand function generator on ebay,
and it's just developed a fatal flaw. It's a Philips (now Fluke)
PM5134, an impressive instrument originally costing a couple of
thousand I think, and which I got for a couple of hundred. The DC
Offset is broken. Not only does it now get nowhere near zero, but the
error is so large that it distorts all waveforms above a small
amplitude.

Use some logic before you start testing components.

It is a function generator, there is an output jack, that is where
you start.

We know that the offset must be added after the function signal, which
is probably coupled to the output through a capacitor.

Between the output of the inner function generator and output jack the
offset voltage must be added.

So if you follow the signal backwards from the output you should find
the offset voltage connection, and then a capacitor, and then the
function generator.

You might be able to disconnect the offset connection, so the function
output is coupled to the output jack without any offset voltage,
through a capacitor which is very likely already present.
Test that capacitor.
Then you can either repair the built-in offset circuit or build an
external one, if you need one.

I hope I have given you some inspiration for some fault finding with
some thought behind, so you don't have to start testing components
aimlessly.


--
Roger J.

 

[Kevin Carney]

Sounds like a bad transistor / IC in the power output stage.

--
change .combo to .com for correct email

***************************************************
"We ought always to know precisely why a given job
is done in a particular way, and why it is done at
all, and why it can't be done more efficiently,
if it must be done at all."-- T.J.Watson

***************************************************

"Roger Johansson" <no-email@home.se> wrote in message
news:8l9ghv4lofub6db216oqoqa1i1h1vkn8kn@4ax.com...

Terry Pinnell <terrypinDELETE@dial.pipexTHIS.com> wrote:

A year or so ago, I bought a second-hand function generator on ebay,
and it's just developed a fatal flaw. It's a Philips (now Fluke)
PM5134, an impressive instrument originally costing a couple of
thousand I think, and which I got for a couple of hundred. The DC
Offset is broken. Not only does it now get nowhere near zero, but the
error is so large that it distorts all waveforms above a small
amplitude.

Use some logic before you start testing components.

It is a function generator, there is an output jack, that is where
you start.

We know that the offset must be added after the function signal, which
is probably coupled to the output through a capacitor.

Between the output of the inner function generator and output jack the
offset voltage must be added.

So if you follow the signal backwards from the output you should find
the offset voltage connection, and then a capacitor, and then the
function generator.

You might be able to disconnect the offset connection, so the function
output is coupled to the output jack without any offset voltage,
through a capacitor which is very likely already present.
Test that capacitor.
Then you can either repair the built-in offset circuit or build an
external one, if you need one.

I hope I have given you some inspiration for some fault finding with
some thought behind, so you don't have to start testing components
aimlessly.


--
Roger J.

 

[Roger Johansson]

Do you think they are moving the whole power output stage (or the
function generator circuit) up and down in voltage level? I don't
think that sounds reasonable.

Or maybe it is reasonable. The function generator itself could be at a
fixed voltage level, and a capacitor couples the signal to a power
stage which can be moved up and down a few volts.

That would allow for a low impedance output with no need for a big
capacitor, and the offset voltage would be equally buffered as it is
implemented together with the power amplifier.

There could be a 50 Ohm resistor directly behind that output jack, or
a switch to select beteen different output impedances.

--
Roger J.

 

[Terry Pinnell]

Roger Johansson <no-email@home.se> wrote:

Do you think they are moving the whole power output stage (or the
function generator circuit) up and down in voltage level? I don't
think that sounds reasonable.

Or maybe it is reasonable. The function generator itself could be at a
fixed voltage level, and a capacitor couples the signal to a power
stage which can be moved up and down a few volts.

That would allow for a low impedance output with no need for a big
capacitor, and the offset voltage would be equally buffered as it is
implemented together with the power amplifier.

There could be a 50 Ohm resistor directly behind that output jack, or
a switch to select beteen different output impedances.

Roger, Kevin, John: Unless my PM5134 is truly deceased, kicked the
bucket, passed away, gone to a better place, etc, etc, Tony Williams
will perhaps soon be able to reveal all. Tony - a star - has very
kindly offered to have a go at diagnosing and fixing this beast. I
posted it 8 hours ago, so, God and the UK Post Office ParcelForce
service permitting, all 9 kg of it (including manual and schematics)
should reach him shortly.

I'm sure he'll be happy to post here to tell you what the problem was,
so you'll probably know before I do. (Hope it's good news!) I'm flying
to Austria for a holiday in early hours of Saturday, returning 2nd
Aug. My wife, Janet, is zealous in keeping me out of Internet Cafes,
so I'll be incommunicado for a fortnight (USA=2 weeks).

--
Terry Pinnell
Hobbyist, West Sussex, UK

 

[Winfield Hill]

Terry wrote...

Roger, Kevin, John: Unless my PM5134 is truly deceased, kicked the
bucket, passed away, gone to a better place, etc, etc, Tony Williams
will perhaps soon be able to reveal all. Tony - a star - has very
kindly offered to have a go at diagnosing and fixing this beast. I
posted it 8 hours ago, so, God and the UK Post Office ParcelForce
service permitting, all 9 kg of it (including manual and schematics)
should reach him shortly.

I'm sure he'll be happy to post here to tell you what the problem was,
so you'll probably know before I do. (Hope it's good news!) I'm flying
to Austria for a holiday in early hours of Saturday, returning 2nd
Aug. My wife, Janet, is zealous in keeping me out of Internet Cafes,
so I'll be incommunicado for a fortnight (USA=2 weeks).

Lessee here, you fly to Austria for holiday, leaving our poor
Tony Williams to slave away on your broken instrument? :&gt

Thanks,
- Win

 

[Tony Williams]

In article <7iqghvspmfmmgccp9be9sanujptqtj489h@4ax.com>,
Terry Pinnell <terrypinDELETE@dial.pipexTHIS.com> wrote:

Roger, Kevin, John: Unless my PM5134 is truly deceased, kicked the
bucket, passed away, gone to a better place, etc, etc, Tony Williams
will perhaps soon be able to reveal all. Tony - a star - has very
kindly offered to have a go at diagnosing and fixing this beast. I
posted it 8 hours ago, so, God and the UK Post Office ParcelForce
service permitting, all 9 kg of it (including manual and schematics)
should reach him shortly.

It arrived this morning... only just, because half the address
label was missing and the PO had to make a wild guess.

I'm flying to Austria for a holiday in early hours of Saturday,
returning 2nd Aug.

The cheeky little tinker! Drops it in the post and then
legs it off abroad for a fortnight.

It's up on the bench now........

The two 20-0-20v supplies and the 5v are ok. Although the two
3-terminal regulators on the main 20-0-20v supply (bolted to the
sidewall) run quite hot.

The output of the final power amp is sitting up at +11.7v, but
that's not unexpected when it has an input of +3.3v.

The 1:1 buffer before that looks to be the perp, it has an input of
0.066v and an output of +3.3v. It is down on the under-board, packed
up against the f.panel, and partly underneath a 2-gang pot... damn
near impossible to reach with even the finest probes.

+20v
|
\
/2k37
\
| 140R
+---------/\/\------+
310R |/e |
+--/\/\--|pnp |
| |\c |
| | 12v |
| 0v--||--+---|<|--- -20V |
Vin ->-+ zeners +--> Vout
66mV | 0v--||--+---|>|--- +20v | 3.3v
| | 12v |
| 310R |/c <----- 7.8v |
+--/\/\--|npn |
/ |\e <----- 5.8v |
61mV______/ +---------/\/\------+
| 140R
\
/2k37
\
|
-20v

I can only reach around the npn for measurement.
Voltage results shown alongside each pin.

That 5.8v emitter voltage looks seriously suss, but the resultant
3.3v for Vout suggests that the pnp might be ok (touch wood, because
changing that one would be a full dismantle of everything).

So I'm going to go off and see if that npn can be changed.

--
Tony Williams.

 

[Fred Bloggs]

Tony Williams wrote:

In article <7iqghvspmfmmgccp9be9sanujptqtj489h@4ax.com>,
Terry Pinnell <terrypinDELETE@dial.pipexTHIS.com> wrote:


Roger, Kevin, John: Unless my PM5134 is truly deceased, kicked the
bucket, passed away, gone to a better place, etc, etc, Tony Williams
will perhaps soon be able to reveal all. Tony - a star - has very
kindly offered to have a go at diagnosing and fixing this beast. I
posted it 8 hours ago, so, God and the UK Post Office ParcelForce
service permitting, all 9 kg of it (including manual and schematics)
should reach him shortly.


It arrived this morning... only just, because half the address
label was missing and the PO had to make a wild guess.


I'm flying to Austria for a holiday in early hours of Saturday,
returning 2nd Aug.


The cheeky little tinker! Drops it in the post and then
legs it off abroad for a fortnight.

It's up on the bench now........

The two 20-0-20v supplies and the 5v are ok. Although the two
3-terminal regulators on the main 20-0-20v supply (bolted to the
sidewall) run quite hot.

The output of the final power amp is sitting up at +11.7v, but
that's not unexpected when it has an input of +3.3v.

The 1:1 buffer before that looks to be the perp, it has an input of
0.066v and an output of +3.3v. It is down on the under-board, packed
up against the f.panel, and partly underneath a 2-gang pot... damn
near impossible to reach with even the finest probes.

+20v
|
\
/2k37
\
| 140R
+---------/\/\------+
310R |/e |
+--/\/\--|pnp |
| |\c |
| | 12v |
| 0v--||--+---|<|--- -20V |
Vin ->-+ zeners +--> Vout
66mV | 0v--||--+---|>|--- +20v | 3.3v
| | 12v |
| 310R |/c <----- 7.8v |
+--/\/\--|npn |
/ |\e <----- 5.8v |
61mV______/ +---------/\/\------+
| 140R
\
/2k37
\
|
-20v

I can only reach around the npn for measurement.
Voltage results shown alongside each pin.

That 5.8v emitter voltage looks seriously suss, but the resultant
3.3v for Vout suggests that the pnp might be ok (touch wood, because
changing that one would be a full dismantle of everything).

So I'm going to go off and see if that npn can be changed.

Good luck- and a very odd circuit- looks like some kind of 75 ohm driver.

 

[Tony Williams]

In article <3F195CFE.20109@nospam.com>,
Fred Bloggs <nospam@nospam.com> wrote:

+20v
|
\
/2k37
\
| 140R
+---------/\/\------+
310R |/e |
+--/\/\--|pnp |
| |\c |
| | 12v |
| 0v--||--+---|<|--- -20V |
Vin ->-+ zeners +--> Vout
66mV | 0v--||--+---|>|--- +20v | 3.3v
| | 12v |
| 310R |/c <----- 7.8v |
+--/\/\--|npn |
/ |\e <----- 5.8v |
61mV______/ +---------/\/\------+
| 140R
\
/2k37
\
|
-20v
So I'm going to go off and see if that npn can be changed.

Good luck-

Tnx. A new transistor seems to have done the trick,
the back end all works properly now.

and a very odd circuit- looks like some kind of 75 ohm
driver.

I've probably just drawn it awkwardly. It looks quite a
neat way of doing a 1:1 buffer to me. The two Vbe's are
subtracted-out, and the zeners clip the signal to 10Vpk.

--
Tony Williams.

 

[Paul Burridge]

On Sat, 19 Jul 2003 22:59:07 +0100, Tony Williams
<tonyw@ledelec.demon.co.uk> wrote:

Tnx. A new transistor seems to have done the trick,
the back end all works properly now.

Well done, Tony. You're a shining example of how Usenet *ought* to be!

 

[John Crighton]

On Sun, 20 Jul 2003 00:08:24 +0100, Paul Burridge
<redscar@waitrose.notformail.com> wrote:

On Sat, 19 Jul 2003 22:59:07 +0100, Tony Williams
tonyw@ledelec.demon.co.uk> wrote:

Tnx. A new transistor seems to have done the trick,
the back end all works properly now.

Well done, Tony. You're a shining example of how Usenet *ought* to be!

I'll second that. Well done, Tony!
Regards,
John Crighton
Sydney

 

[Tony Williams]

In article <3f19e264.2420410@News.CIS.DFN.DE>,
John Crighton <john_c@tpg.com.au> wrote:
[and PB]

I'll second that. Well done, Tony!

I think there was a lot of luck in it..... just one stage upstream
of the power amp, a clearly identifiable fault, and in just one
component (and even the accessible one of the pair). It could have
been far worse. And I'm still uncomfortable with those hot-running
three-terminal regulators.

There are a lot of tired-looking components and solder joints
in there, and at least one of the voltages is off the range of
its trimmer, so TP may have to do some preventative maintenance.

It's a wide-ranging (0.1Hz to 20MHz) Triangle/Sine/Square
function generator, done mainly in discrete transistors, and
possibly representative of the top-end design of the day. It
is very interesting to work through the circuitry, trying to
see why he did this or that, etc. Perhaps Terry can scan parts
of it and post them for discussion (when he comes back from
his jaunt around Austria of course).

--
Tony Williams.

 

[Tony Williams]

On 20 Jul, Tony Williams <tonyw@ledelec.demon.co.uk> wrote:

And I'm still uncomfortable with those hot-running
three-terminal regulators.

Hmph... the transformer is set to the 220v tapping,
and UK mains is supposed to be 230v nominal, but
much of it still runs at the old 240v. Ours is at
242v this morning. Perhaps that is why things run
a little warm in there. I'll leave the tap-change
for Terry to decide (and do).

--
Tony Williams.

 

[Andre]

Paul Burridge <redscar@waitrose.notformail.com> wrote in message news:<rnjjhvgmqudicu7p2fcahkker5ucpnokpa@4ax.com>...

On Sat, 19 Jul 2003 22:59:07 +0100, Tony Williams
tonyw@ledelec.demon.co.uk> wrote:

Tnx. A new transistor seems to have done the trick,
the back end all works properly now.

Well done, Tony. You're a shining example of how Usenet *ought* to be!

Yup . I'll second that

-A

 

[Winfield Hill]

Tony Williams <tonyw@ledelec.demon.co.uk> wrote ...

Terry Pinnell terrypinDELETE@dial.pipexTHIS.com> wrote:

Roger, Kevin, John: Unless my PM5134 is truly deceased, kicked the
bucket, passed away, gone to a better place, etc, etc, Tony Williams
will perhaps soon be able to reveal all. Tony - a star - has very
kindly offered to have a go at diagnosing and fixing this beast. I
posted it 8 hours ago, so, God and the UK Post Office ParcelForce
service permitting, all 9 kg of it (including manual and schematics)
should reach him shortly.

It's up on the bench now........

The 1:1 buffer before that looks to be the perp, it has an input of
0.066v and an output of +3.3v. It is down on the under-board, packed
up against the f.panel, and partly underneath a 2-gang pot... damn
near impossible to reach with even the finest probes.

+20v
|
\
/2k37
\
| 140R
+---------/\/\------+
310R |/e |
+--/\/\--| pnp |
| |\c |
| | 12v |
| 0v--||--+---|<|--- -20V |
Vin ->-+ zeners +--> Vout
66mV | 0v--||--+---|>|--- +20v | 3.3v
| | 12v |
| 310R |/c <----- 7.8v |
+--/\/\--| npn |
/ |\e <----- 5.8v |
61mV______/ +---------/\/\------+
| 140R
\
/2k37
\
|
-20v

I can only reach around the npn for measurement.
Voltage results shown alongside each pin.

That 5.8v emitter voltage looks seriously suss, but the resultant
3.3v for Vout suggests that the pnp might be ok (touch wood, because
changing that one would be a full dismantle of everything).

So I'm going to go off and see if that npn can be changed.

Better see if that pnp can be changed. It's supposed to pull the node
in question down and isn't doing so. The npn can only pull up, right?