29 Jan
2009
29 Jan
'09
8:59 a.m.
Hi,
I have a couple of uPDP-11 boxes (BA23) with failed power supplies.
Being no expert on repairing switch mode power supplies (and to be frank
slighly scared of trying), is there any reason why I shouldn't adapt a
ATX power supply from a PC to provide the required voltages?
Most ATX power supplies offer pretty meaty 5V and 12V feeds plus a
fairly small -12V (~0.3A) feed.
Will I need to generate a LTC/BEVENT L signal if I go down this route?
Finally, does anyone know of any good guides on repairing switch modes?
Whilst not keen, I guess I should bite the bullet (hopefully not
literally) and learn how to repair them.
Thanks,
Toby
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
29 Jan
29 Jan
11:49 a.m.
Finally, does anyone know of any good guides on
repairing switch modes?
Whilst not keen, I guess I should bite the bullet (hopefully not
literally) and learn how to repair them.
I've repaired SMPUs... The first one terrified me, now I am still
cautious, but I will do them.
Until you get some expernince, don't even think of repairing a PSU
without a scheamtic. There are several difference circuits, the control
circuitryt may be on the mains side or the output side of the 'isolation
barrier' for example, and not realising whcih is is a recipe for giving
yourself a possibly fatal shock amd more importatnly damaging the
machine. (Yes I ahve got that the right way round. There are many more
hackers than classic computer out there...)
There used to be a 'repair guide to SMPSUs' (or some similar title) as
part of the Usenet sci.electronics.repair FAQ. That is worth trying to
find and read.
Also, if you're in the UK, try reading 'Television' magazine from around
1985 onwards (or maybe a bit before that). Many TVs of that period used
standard-ish SMPSU circuits, and the repair methods are much the same for
computer supplies.
In general, stanr by looking at the fuse. If it has blown violently
(shatered or blackened glass), suspect a major short on the mains side of
the PSU. If it's a rectifier diode, then most likely it was not damaged
by anything else, and replacing it will get the supply working again. If
the chopper transistor has shorted (very common, alas), then it may well
have been damafgd by somthing else. Maybe a shorted rectifier or
capacitor on the secondary side (but the protection circuit _should_ cope
with that), maybe a problem with the snubber network, maybe shorted turns
in the chopper transformer (A replacement is avialalbe 'locally' as the
manuals say, where 'locally' is either a misprint for luckily, or the
name of an oscure Welsh village :-)).
If the fuse is OK, and the supply is 'tweeting' (trying to start up, then
shutting down, and repeating), then the problem could be a short on the
scondary side, it may be a fault in the voltage regulation loop (so that
the crowbar is tripping, shorting out the supply and shutting things
down), or maybe an open-circuit or high ESR capacitor. The last is very
common, some people who go in for 'lucky dip servicing' just test all the
elctrolytics in the SMPSU with an ESR meter and replace any that seem a
little high.
If the fuse is OK and the supply seeems dead, then probably the startup
circuit has failed. In smaller supplies, this is often a high-value
resistor (220K or so) from the +ve side of the mains smoothing capacitor
into the depths of the chopper circuit. It oftne fails open-circuit. If
you can find it, test it.
Beware of the last fault. Some SMPSUs don't have bleeder resistors on the
mains smoothing capacitor, relying on the chopper circuit to discharge
them. If this isn't running for any reason, the capacitor stays charged
and will zap you. Now how do you think UI found that out :-)
-tony
12:11 p.m.
On 29 Jan 2009 at 19:49, Tony Duell wrote:
Beware of the last fault. Some SMPSUs don't have
bleeder resistors on the
mains smoothing capacitor, relying on the chopper circuit to discharge
them. If this isn't running for any reason, the capacitor stays charged
and will zap you. Now how do you think UI found that out :-)
Although the basic principle is simple, SMPSU design to me is still a
black art and it seems that the practitioners of it are all older
than 50. One of the better websites on the subject is:
http://www.smpstech.com/index.htm
FWIW,
Chuck
30 Jan
30 Jan
10:52 a.m.
On 29 Jan 2009 at 19:49, Tony Duell wrote:
The design _is_ something of a black art (in particular the design of the
transformer -- the exact characteristics of that are critical to the
design). But repairing an SMPSU that has worked (and in th case of a DEC
Q-bus box, has been made many time,s and has worked well for years in
general) is a lot easier. You know the basic deisgn ins sound, you know
that particular transformer can work in that circuit, and so on.
Beware of the last fault. Some SMPSUs don't
have bleeder resistors on the
mains smoothing capacitor, relying on the chopper circuit to discharge
them. If this isn't running for any reason, the capacitor stays charged
and will zap you. Now how do you think UI found that out :-)
Although the basic principle is simple, SMPSU design to me is still a
black art and it seems that the practitioners of it are all older than 50. One of the better websites on the subject
is:
http://www.smpstech.com/index.htm
Other references are the sevice manauls for some early-ish devices that
used such supplies (the HP7245 printer/plotter manual springs to mind as
containing an explanation of a '2 stage' SMPSU). And the second editon of
The Art of Electronics contains a basic description of the Tandy 2000
PSU, complete with a schematic (and an insulting comment about IBM and
Apple who refused to provide scheamtics!).
-tony
11:47 a.m.
On 30 Jan 2009 at 18:52, Tony Duell wrote:
The design _is_ something of a black art (in
particular the design of the
transformer -- the exact characteristics of that are critical to the
design). But repairing an SMPSU that has worked (and in th case of a DEC
Q-bus box, has been made many time,s and has worked well for years in
general) is a lot easier. You know the basic deisgn ins sound, you know
that particular transformer can work in that circuit, and so on.
While repair is interesting in its own right, it's the design that
really shows the mind (or lack thereof) of the engineer. For
example, it never occurred to me that putting an EMI line filter in
front of a SMPSU, while it may not disable, may compromise its
operation.
How many engineers have simply spec-ed an off-the-shelf PSU and stuck
it after one of those dandy little fuse-receptacle-power-switch-EMI
filter boxes without thinking to ask the PSU manufacturer if it was
advisable to do so? It certainly wouldn't have occurred to me.
Cheers,
Chuck
1:41 p.m.
On Friday 30 January 2009 02:47:56 pm Chuck Guzis wrote:
On 30 Jan 2009 at 18:52, Tony Duell wrote:
Eh? Why would it do that?
The design _is_ something of a black art (in
particular the design of the
transformer -- the exact characteristics of that are critical to the
design). But repairing an SMPSU that has worked (and in th case of a DEC
Q-bus box, has been made many time,s and has worked well for years in
general) is a lot easier. You know the basic deisgn ins sound, you know
that particular transformer can work in that circuit, and so on.
While repair is interesting in its own right, it's the design that
really shows the mind (or lack thereof) of the engineer. For
example, it never occurred to me that putting an EMI line filter in
front of a SMPSU, while it may not disable, may compromise its
operation. How many engineers have simply spec-ed an
off-the-shelf PSU and stuck
it after one of those dandy little fuse-receptacle-power-switch-EMI
filter boxes without thinking to ask the PSU manufacturer if it was
advisable to do so? It certainly wouldn't have occurred to me.
Me neither...
--
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space, a critter that can
be killed but can't be tamed. --Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James
M Dakin
2:31 p.m.
On 30 Jan 2009 at 16:41, Roy J. Tellason wrote:
Eh? Why would it do that?
Read about it:
http://www.smpstech.com/filter00.htm
Interestingly, the real-world example cited is the addition of an
input filter to a PDP 11/40...
Cheers,
?huck
3:05 p.m.
On Friday 30 January 2009 05:31:20 pm Chuck Guzis wrote:
On 30 Jan 2009 at 16:41, Roy J. Tellason wrote:
Who'd a thunk it...?
Eh? Why would it do that?
Read about it:
http://www.smpstech.com/filter00.htm Interestingly, the real-world example cited is the
addition of an
input filter to a PDP 11/40...
I read that as him trying to use that machine to simulate an aspect of the
problem, or similar.
--
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space, a critter that can
be killed but can't be tamed. --Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James
M Dakin
4:13 p.m.
Chuck Guzis wrote:
On 30 Jan 2009 at 16:41, Roy J. Tellason wrote:
I believe in that discussion he's referring to objects in their pure form: when
he says SMPS he means just the DC-to-DC converter; not what we colloquially
refer to as a SMPS, which is really a conventional AC-DC converter (rectifier &
filter) followed by a switch-mode DC-DC converter.
As I would understand it, the problem with adding the EMI filter to the front
of the SM DC-DC converter is that the converter actually draws current from the
DC source at a frequency corresponding to the switch-mode frequency, so the
current through an input EMI filter is nothing like DC, and you can get AC
interaction between the filter and the DC-DC converter. The EMI filter becomes
the resonant circuit, and negative resistance characteristics appearing at the
the input of the DC-DC converter turn the combination into an oscillator.
I believe his discussion is intended to be more of pertinence to DC
distribution-bus systems such as in aircraft or telecom (note he refers in a
couple of places to 28V distribution).
I don't believe he is discussing the situation of adding an EMI filter to the
front of a typical 120/240VAC input SMPS. In such supplies the 'lytic filter
caps of the AC-DC converter at the beginning of the supply reduce the
current-variation frequency to 50/100/60/120Hz (unless there is inadequate
filtering), which is not a problem for the EMI filters.
(Of course, these supplies generally do include built-in EMI filters at the front).
Eh? Why would it do that?
Read about it:
http://www.smpstech.com/filter00.htm
4:31 p.m.
On 30 Jan 2009 at 16:13, Brent Hilpert wrote:
I believe in that discussion he's referring to
objects in their pure form: when
he says SMPS he means just the DC-to-DC converter; not what we colloquially
refer to as a SMPS, which is really a conventional AC-DC converter (rectifier &
filter) followed by a switch-mode DC-DC converter.
As I would understand it, the problem with adding the EMI filter to the front
of the SM DC-DC converter is that the converter actually draws current from the
DC source at a frequency corresponding to the switch-mode frequency, so the
current through an input EMI filter is nothing like DC, and you can get AC
interaction between the filter and the DC-DC converter. The EMI filter becomes
the resonant circuit, and negative resistance characteristics appearing at the
the input of the DC-DC converter turn the combination into an oscillator.
Thanks for the elaboration! I hadn't read the whole thing yet, but
was mildly surprised at the lede. Your explanation makes sense.
Cheers,
Chuck
31 Jan
31 Jan
12:09 a.m.
Chuck Guzis wrote:
On 30 Jan 2009 at 16:13, Brent Hilpert wrote:
Thanks for the elaboration! I hadn't read the whole thing yet, but
was mildly surprised at the lede. Your explanation makes sense.
Hmmm.. thinking about it a bit more, maybe I better backtrack a bit:
I believe
in that discussion he's referring to objects in their pure form: when
he says
SMPS he means just the DC-to-DC converter; not what we colloquially
refer to as a SMPS, which is really a conventional AC-DC converter (rectifier &
filter) followed by a switch-mode DC-DC converter.
As I would understand it, the problem with adding the EMI filter to the front
of the SM DC-DC converter is that the converter actually draws current from the
DC source at a frequency corresponding to the switch-mode frequency, so the
current through an input EMI filter is nothing like DC, and you can get AC
interaction between the filter and the DC-DC converter. The EMI filter becomes
the resonant circuit, and negative resistance characteristics appearing at the
the input of the DC-DC converter turn the combination into an oscillator. > I don't believe he is discussing the
situation of adding an EMI filter to the
> front of a typical 120/240VAC input SMPS. In such supplies the 'lytic filter
> caps of the AC-DC converter at the beginning of the supply reduce the
> current-variation frequency to 50/100/60/120Hz (unless there is inadequate
> filtering),
When the rectifiers of the AC-DC converter turn on (conduct) at the peaks of
the 50/60Hz mains sine wave, there will be a switch-mode-frequency component in
the current drawn from the mains, so there may still be an issue for an AC
mains EMI filter. How long those bursts are depends on the conduction angle of
the rectifiers, as determined by the final load demand, but it will be many
cycles at the switch-mode f.
Perhaps this is (one reason) why these supplies have built-in EMI filters on
the AC input, so that that filter can be tailored to the rest of the SM design.
Nonetheless, I suspect that by the time you get to the front of that internal
EMI filter (to the IEC connector or outside the power supply box, so to speak),
there is no longer ((in modern supplies), much of) a problem, as this (in part)
is what the internal EMI filter is there to deal with.
I wonder if this is one reason that early SMPS's, when perhaps such issues as
this were not dealt with as well, were prone to seemingly arbitrary self-destruction.
(.. All of which serves to reinforce your original point: that the functioning
of these things involves a lot of issues beyond the basic principles as
commonly presented.)
10:51 a.m.
On 30 Jan 2009 at 16:41, Roy J. Tellason wrote:
Which would be curious if true, since no standard PDP11/40 used an SMPSU.
Switching _regulators_, sure, but the PSU starts with a large
mains-frequency transofrmer, which is not going to have much of an HF
response.
Howev,er I read the article as that the author had run some simulation
software on an 11/40 to study the preformace of another PSU. And that
makes a lot mroe sense.
-tony
Eh? Why would it do that?
Read about it:
http://www.smpstech.com/filter00.htm
Interestingly, the real-world example cited is the addition of an=20
input filter to a PDP 11/40... 
1 Feb
1 Feb
3:32 a.m.
On Sat, 2009-01-31 at 18:51 +0000, Tony Duell wrote:
Howev,er I read the article as that the author had run
some simulation
software on an 11/40 to study the preformace of another PSU. And that
makes a lot mroe sense.
That's what I read too, and the whole article made sense.
Simulation == all bets are off. It's never going to tell you anything
meaningful.
Gordon
7:34 a.m.
Gordon JC Pearce MM3YEQ wrote:
Simulation == all bets are off. It's never going
to tell you anything
meaningful.
That's a bit of a sweeping statement! I agree badly done simulation is
worse than useless, but I have to use simulation all the time in my job
(power station planning). Some of the things I frequently have to tell
both trainees and clients are worth mentioning here, though:
1. The model is not the system. "The model predicts it" is not an
explanation for anything. Any prediction of the model has to represent
a physical process on the system you are studying, and you need to
understand what that process is and why it happens [1]. (Or, for that
matter, why it doesn't behave as the model predicts)
2. No one model is good for everything. You need different models for
different studies. Before you use a model, it is important to
understand what it is for, and what it can do. This often involves
quite an in-depth look at how the simulation software actually uses the
data. [2]
3. You can't make a model more accurate by throwing a bigger computer
at it. (So many clients don't appreciate this!) A model stands or
falls by the data that goes in it. It doesn't matter how sophisticated
the simulation software, if you haven't got good data you can't do an
accurate model. GIGO.
Take those three things into account, and simulation is a useful tool.
But it's no more than a tool, and some of the best simulations are very
simple models that run on that ubiquitous platform, the human brain...
Philip.
[1] I remember long arguments with the suppliers of some simulation
software on this subject. They insisted that a second harmonic
oscillation should die away so quickly they didn't need to model it. I
maintained that it was a result of dc flowing in the stator of an ac
generator (the Hammond Organ effect), and should die away no faster than
the dc that caused it.
[2] Which is why the engineer at the network company (I'll not name
them) couldn't reproduce my results. She was trying to cheat by putting
her calculated time values of impedance in the simulation software's
initial value, and didn't realise what else it would affect. To give
her the credit that's her due, she realised what was wrong as soon as I
pointed out how the software calculated one of the impedances. But I
shouldn't have had to - and it was just luck we both had the same
software...
1:23 p.m.
Simulation == all bets are off. It's never going
to tell you anything
meaningful.
I wouldn't go that far, for all I am one of the people least likely to
use or trust a simulation. I would say that a simulation is only as good
as the model and the data, and that in most cases you haven't much of a clue
about how either of those relates to reality.
For example, a circuit similation may only be meanigful if you correctly
handle the inductance of all the wires and the cpaactiance between them.
Something that you might well not know accurately enough!.
If curse correctly, simulation can be useful. IMHO one of the frw correct
ways (if not the only way) to use simulation is :
1) You have a real physical system that you do experiments on and collect
real data
2) You simulate the same system under the same conditions and simulate
the same experiements on that system
3) If the results agree, then it is likely (although not certain) that
the simulation is an accurate description of the system and that tghe
theory you used to produce the cimulation also applies to the real system
But if they results don't agree then either you have experimental
measurement errors, a bug in the program, or more interestingly the
theory doesn't apply to the real system, perhpas there is some other
effect you've neglected. In which case you need to do more investigations.
-tony
3:13 p.m.
On Sunday 01 February 2009, Gordon JC Pearce MM3YEQ wrote:
Simulation == all bets are off. It's never going
to tell you
anything meaningful.
If this was true, many people like me wouldn't have a job.
There's many scientific disciplines where simulation is much more
practical than physical experimentation, to help hypothesis.
The biggest example is funded by the DOE to do nuclear detonation
simulations.
Pat
--
Purdue University Research Computing --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org
4:09 p.m.
On Sun, 2009-02-01 at 18:13 -0500, Patrick Finnegan wrote:
On Sunday 01 February 2009, Gordon JC Pearce MM3YEQ
wrote:
Yes, because it's hard to test nuclear explosions on your bench - twice,
anyway.
I've lost count of the number of people I've spoken to in ##electronics
and on forums who have got hopelessly lost and confused by trying to
simulate simple electronic circuits. Just build the damn thing and test
it, you'll learn far more! But no, they don't want to do that because
apparently "it's a waste of time and money".
Gordon
Simulation == all bets are off. It's never
going to tell you
anything meaningful.
If this was true, many people like me wouldn't have a job.
There's many scientific disciplines where simulation is much more
practical than physical experimentation, to help hypothesis.
The biggest example is funded by the DOE to do nuclear detonation
simulations.
4:27 p.m.
On 2 Feb 2009 at 0:09, Gordon JC Pearce MM3YEQ wrote:
I've lost count of the number of people I've
spoken to in ##electronics
and on forums who have got hopelessly lost and confused by trying to
simulate simple electronic circuits. Just build the damn thing and test
it, you'll learn far more! But no, they don't want to do that because
apparently "it's a waste of time and money".
Simulation is an essential part of working with programmable (i.e.
SystemC, Verilog or VHDL) devices. Being able to write a good
simulation testbench is probably just as valuable as being able to
write the original programming for the device.
Eventually, you have to burn your design into the d*mned thing and do
some real-world testing, but simulation today is just as important as
knowing how to use a scope or logic analyzer.
Cheers,
Chuck
2 Feb
2 Feb
1:05 a.m.
On Sun, 2009-02-01 at 16:27 -0800, Chuck Guzis wrote:
On 2 Feb 2009 at 0:09, Gordon JC Pearce MM3YEQ wrote:
Again, that's all very well if you're doing something sufficiently
complex and well-defined. If you are investigating the effect of
putting what is basically a tuned circuit across the supply to a
non-linear load, where a) it's simple enough to build, and b) it's not
clearly defined enough to model, then you're far better just warming up
the soldering iron and breaking out the 'scope.
Likewise, a lot of people seem to be taught that the only way to
investigate a circuit's behaviour is to simulate it. When someone shows
me a simulator that can accurately reproduce the behaviour of a simple
twin-T filter under all possible conditions then I'll start to warm to
them. A circuit that doesn't work properly doesn't sit there Just Plain
Doing Nothing, or producing an odd error message.
Gordon
I've lost count of the number of people
I've spoken to in ##electronics
and on forums who have got hopelessly lost and confused by trying to
simulate simple electronic circuits. Just build the damn thing and test
it, you'll learn far more! But no, they don't want to do that because
apparently "it's a waste of time and money".
Simulation is an essential part of working with programmable (i.e.
SystemC, Verilog or VHDL) devices. Being able to write a good
simulation testbench is probably just as valuable as being able to
write the original programming for the device. 
12:57 p.m.
Again, that's all very well if you're doing
something sufficiently
complex and well-defined. If you are investigating the effect of
putting what is basically a tuned circuit across the supply to a
non-linear load, where a) it's simple enough to build, and b) it's not
clearly defined enough to model, then you're far better just warming up
the soldering iron and breaking out the 'scope.
Until the boss sees that you screwed around in the lab for an hour or
two getting all the parts and building the test circuit, rather than
spend fifteen minutes at your workstation in your cube, just to get
the same results.
--
Will
2:36 p.m.
On Mon, 2 Feb 2009, William Donzelli wrote:
Until the boss sees that you screwed around in the lab for fifteen minutes
without a demonstrable product, when you could have been spending an hour
or two at your workstation in your cubicle, just to get the same
results.
?
Again,
that's all very well if you're doing something sufficiently
complex and well-defined. If you are investigating the effect of
putting what is basically a tuned circuit across the supply to a
non-linear load, where a) it's simple enough to build, and b) it's not
clearly defined enough to model, then you're far better just warming up
the soldering iron and breaking out the 'scope.
Until the boss sees that you
screwed around in the lab for an hour or
two getting all the parts and building the test circuit, rather than
spend fifteen minutes at your workstation in your cube, just to get
the same results. 
4 Feb
4 Feb
11:59 a.m.
On Feb 2, 2009, at 3:57 PM, William Donzelli wrote:
Heh...I think you have those time estimates reversed.
-Dave
--
Dave McGuire
Port Charlotte, FL
Again,
that's all very well if you're doing something sufficiently
complex and well-defined. If you are investigating the effect of
putting what is basically a tuned circuit across the supply to a
non-linear load, where a) it's simple enough to build, and b) it's
not
clearly defined enough to model, then you're far better just
warming up
the soldering iron and breaking out the 'scope.
Until the boss sees that you screwed around in the lab for an hour or
two getting all the parts and building the test circuit, rather than
spend fifteen minutes at your workstation in your cube, just to get
the same results.
1:27 p.m.
On Feb 4, 2009, at 3:19 PM, William Donzelli wrote:
Ahh, suitly companies. Nope, I avoid them like the plague. *I*
have a well-stocked lab.
-Dave
--
Dave McGuire
Port Charlotte, FL
Heh...I think
you have those time estimates reversed.
You have not dealt with a modern, barely stocked, corporate lab
lately, have you?
"Hey, I need a few 100 ohm resistors. Have any back there?"
"Nope, try Digikey."
5 Feb
5 Feb
4:04 p.m.
On Wednesday 04 February 2009 04:27:42 pm Dave McGuire wrote:
On Feb 4, 2009, at 3:19 PM, William Donzelli wrote:
Ahh, suitly companies. Nope, I avoid them like the plague.
Too bad that's all there seems to be around these parts... :-(
That's why I'm doing a bit of contracting these days. Is it different where
you are?
Heh...I
think you have those time estimates reversed.
You have not dealt with a modern, barely stocked, corporate lab
lately, have you?
"Hey, I need a few 100 ohm resistors. Have any back there?"
"Nope, try Digikey." *I* have a well-stocked lab.
:-)
--
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space, a critter that can
be killed but can't be tamed. --Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James
M Dakin
31 Jan
31 Jan
10:46 a.m.
On 30 Jan 2009 at 18:52, Tony Duell wrote:
Agreed, but the OP was interested in repairing them, I believe.
The design _is_ something of a black art (in
particular the design of the
transformer -- the exact characteristics of that are critical to the
design). But repairing an SMPSU that has worked (and in th case of a DEC
Q-bus box, has been made many time,s and has worked well for years in
general) is a lot easier. You know the basic deisgn ins sound, you know
that particular transformer can work in that circuit, and so on.
While repair is interesting in its own right, it's the design that
really shows the mind (or lack thereof) of the engineer. For example, it never occurred to me that putting an EMI
line filter in
front of a SMPSU, while it may not disable, may compromise its
operation.
Presumably because the SMPSU can draw 'spikes' of current when the
chopper decice turens on. Such spikes have a significant high-frequency
component..
However, I am pretty sure that any supply that had a significant
high-frequency compoent to its input current would not meet the
appropriate regulations in Europe, and could not carry a CE mark.
It would also be a poor design IMHO. You have no idea of the
high-frequency characteristics of the supply mains. While the
low-frewuency impendance is clearly pretty low (think of the short
circuit current :-)), the HF impecnace is going to be a lot higher.
How many engineers have simply spec-ed an
off-the-shelf PSU and stuck
it after one of those dandy little fuse-receptacle-power-switch-EMI
filter boxes without thinking to ask the PSU manufacturer if it was
advisable to do so? It certainly wouldn't have occurred to me.
Err, have you ever tried getting useful information out of manufacturers?
-tony
12:33 p.m.
On 31 Jan 2009 at 18:46, Tony Duell wrote:
Err, have you ever tried getting useful information
out of manufacturers?
Depends on your association. If your firm orders SMPSUs by the
thousand, you tend not to have much trouble getting information--if
the vendor wants to keep the account.
I never had much trouble dealing with Intel Application engineers
when working with the early steppings of the 80186 and 80286. It
didn't hurt that Bill Davidow sat on our board, however...
Cheers,
Chuck
29 Jan
29 Jan
1:48 p.m.
Hi Tony,
Thanks for the excellent SMPSU 101. I'm beginning to wish I hadn't
lounged in bed instead of attending the power electronics course at
university (Monday 9am lectures were always going to be a mistake).
Thanks,
Toby
On Thu, 2009-01-29 at 19:49 +0000, Tony Duell wrote:
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Finally, does
anyone know of any good guides on repairing switch modes?
Whilst not keen, I guess I should bite the bullet (hopefully not
literally) and learn how to repair them.
I've repaired SMPUs... The first one terrified me, now I am still
cautious, but I will do them.
Until you get some expernince, don't even think of repairing a PSU
without a scheamtic. There are several difference circuits, the control
circuitryt may be on the mains side or the output side of the 'isolation
barrier' for example, and not realising whcih is is a recipe for giving
yourself a possibly fatal shock amd more importatnly damaging the
machine. (Yes I ahve got that the right way round. There are many more
hackers than classic computer out there...)
There used to be a 'repair guide to SMPSUs' (or some similar title) as
part of the Usenet sci.electronics.repair FAQ. That is worth trying to
find and read.
Also, if you're in the UK, try reading 'Television' magazine from around
1985 onwards (or maybe a bit before that). Many TVs of that period used
standard-ish SMPSU circuits, and the repair methods are much the same for
computer supplies.
In general, stanr by looking at the fuse. If it has blown violently
(shatered or blackened glass), suspect a major short on the mains side of
the PSU. If it's a rectifier diode, then most likely it was not damaged
by anything else, and replacing it will get the supply working again. If
the chopper transistor has shorted (very common, alas), then it may well
have been damafgd by somthing else. Maybe a shorted rectifier or
capacitor on the secondary side (but the protection circuit _should_ cope
with that), maybe a problem with the snubber network, maybe shorted turns
in the chopper transformer (A replacement is avialalbe 'locally' as the
manuals say, where 'locally' is either a misprint for luckily, or the
name of an oscure Welsh village :-)).
If the fuse is OK, and the supply is 'tweeting' (trying to start up, then
shutting down, and repeating), then the problem could be a short on the
scondary side, it may be a fault in the voltage regulation loop (so that
the crowbar is tripping, shorting out the supply and shutting things
down), or maybe an open-circuit or high ESR capacitor. The last is very
common, some people who go in for 'lucky dip servicing' just test all the
elctrolytics in the SMPSU with an ESR meter and replace any that seem a
little high.
If the fuse is OK and the supply seeems dead, then probably the startup
circuit has failed. In smaller supplies, this is often a high-value
resistor (220K or so) from the +ve side of the mains smoothing capacitor
into the depths of the chopper circuit. It oftne fails open-circuit. If
you can find it, test it.
Beware of the last fault. Some SMPSUs don't have bleeder resistors on the
mains smoothing capacitor, relying on the chopper circuit to discharge
them. If this isn't running for any reason, the capacitor stays charged
and will zap you. Now how do you think UI found that out :-)
-tony
30 Jan
30 Jan
11:04 a.m.
Hi Tony,
Thanks for the excellent SMPSU 101. I'm beginning to wish I hadn't
lounged in bed instead of attending the power electronics course at
university (Monday 9am lectures were always going to be a mistake).
Well, you can take heart from the fact that I've never done a single
electronics course. I am entirely self-taught.
[In fact, had I not neen sufficiently interested to teach myself the
interesting bits of the subjects I was studying for A-level and had
relied on the garbage the teachers were spouting, I'd have eneded up
doing nothing of any use.]
-tony
29 Jan
29 Jan
1:35 p.m.
On 29/01/2009 16:59, Tobias Russell wrote:
I have a couple of uPDP-11 boxes (BA23) with failed
power supplies.
Being no expert on repairing switch mode power supplies (and to be frank
slighly scared of trying), is there any reason why I shouldn't adapt a
ATX power supply from a PC to provide the required voltages?
Most ATX power supplies offer pretty meaty 5V and 12V feeds plus a
fairly small -12V (~0.3A) feed.
Will I need to generate a LTC/BEVENT L signal if I go down this route?
Yes, you will, although (a) it's not hard, and (b) you could simply add
a timer card that does it, like a KWV11. We had this discussion very
recently, over a BA11-V box.
However, that's not all you need. There are two distinct bus signals
that tell the CPU (and potentially other devices) that the power is
good. There's one to say the AC supply to the PSU is good (BPOK) and
another to say the DC is good (BDCOK); these are essential to some
processors for a clean startup and are used to control the CPU shutdown
and power-fail/restart.
Finally, does anyone know of any good guides on
repairing switch modes?
Whilst not keen, I guess I should bite the bullet (hopefully not
literally) and learn how to repair them.
Tony :-)
--
Pete Peter Turnbull
Network Manager
University of York
5579
days inactive
5587
days old
29 comments
12 participants
participants (12)
-
ard＠p850ug1.demon.co.uk -
cclist＠sydex.com -
cisin＠xenosoft.com -
gordonjcp＠gjcp.net -
hilpert＠cs.ubc.ca -
mcguire＠neurotica.com -
pat＠computer-refuge.org -
pete＠dunnington.plus.com -
philip＠axeside.co.uk -
rtellason＠verizon.net -
toby＠coreware.co.uk -
wdonzelli＠gmail.com