The best tool for repairing switching supplies (or monitors, for that matter) - an ESR meter. Out of all of the test gear I've purchased over the years, this one probably gets the most use! They use such a low voltage, you can usually use them in-circuit without having to desolder the component off the board before testing. Only takes a few minutes to go through a power supply and identify the suspect electrolytics.
Ian
On Feb 5, 2014, at 12:05 PM, Dave McGuire <mcguire at neurotica.com> wrote:
On 02/05/2014 04:17 AM, Mark Wickens wrote:
Fans would click for a second and then kick right back off. It
stopped doing that until I removed all the drives. Plugged 'em back
in after opening the PSU to look for obvious faults and found none.
Everything is working fine now...Strange.
I had a hard disk taken out by a power spike. I also had firmware
corruption that required reflashing the EEPROM, but I don't remember if
that was related to the PSU. They draw 30 watts when switched 'off' so
it's definitely worth turning them off when not in use.
Twitching fans might indicate over-current protection kicking in.
If it's a repetitive twitching, it's probably the error amplifier or
the voltage divider feeding it. (those will be difficult to test
individually unless they're discretes, of course)
One other thing I've seen recently with older switching power supplies
is the output capacitors' ESR going up due to age, causing their time
constant to exceed that of the regulation loop...creating, you guessed
it, an oscillator.
I know one thing however - switched mode PSU experts are few and far
between!
This is painfully true. I *design* the damn things and I shy away
from working on them most of the time. Direct AC-driven ones are RIGHT
OUT in my book.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
---
Filter service subscribers can train this email as spam or not-spam here: http://my.email-as.net/spamham/cgi-bin/learn.pl?messageid=EA9D86968EA011E39…
On 02/05/2014 04:17 AM, Mark Wickens wrote:
Fans would click for a second and then kick right back off. It
stopped doing that until I removed all the drives. Plugged 'em back
in after opening the PSU to look for obvious faults and found none.
Everything is working fine now...Strange.
I had a hard disk taken out by a power spike. I also had firmware
corruption that required reflashing the EEPROM, but I don't remember if
that was related to the PSU. They draw 30 watts when switched 'off' so
it's definitely worth turning them off when not in use.
Twitching fans might indicate over-current protection kicking in.
If it's a repetitive twitching, it's probably the error amplifier or
the voltage divider feeding it. (those will be difficult to test
individually unless they're discretes, of course)
One other thing I've seen recently with older switching power supplies
is the output capacitors' ESR going up due to age, causing their time
constant to exceed that of the regulation loop...creating, you guessed
it, an oscillator.
I know one thing however - switched mode PSU experts are few and far
between!
This is painfully true. I *design* the damn things and I shy away
from working on them most of the time. Direct AC-driven ones are RIGHT
OUT in my book.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On 05/02/2014 08:00, Cory Smelosky wrote:
On Wed, 5 Feb 2014, Google wrote:
On 5 Feb 2014, at 05:26, Cory Smelosky <b4 at gewt.net> wrote:
So, my VAXstation 4000/60 shut off and I had difficulty getting it to turn back on. Seems like a not-to-abnormal situation, right? PSU's getting old or overheating protection kicked in? That's not the weird part!
It may have overheated, I do know power switches can get flakey in VS4000s but the symptoms tend to be it won't turn OFF not on't turn ON but I guess if the switch is not contacting correctly it could be that too. the switch isn't, as far as I can tell, directly controlling the current, it is in some kind of latch circuit that unlatches when you flick it to 'OFF'. I think it's an early 'soft-power' implementation.
I moved the UPS further away. Very possible it got bumped and got too close to the UPS.
Full explanation:
http://dectec.info/vaxstation-4000-power-switch-issues-and-cleaning/
Thanks!
Also talk to Mark Wickens, he's had a few issues with 4000/90 PSUs.
When I rebooted the system...it had decided it was suddenly /2015/ and all the licenses expired. I've heard of systems resetting to the past when something happens...but NEVER the future!
That is odd, but Mark W remarked on my blog about one PSU failing and spiking his 4000/90 so badly it fried several parts, it's possible it could have upskittled the TOY clock?
Fans would click for a second and then kick right back off. It stopped doing that until I removed all the drives. Plugged 'em back in after opening the PSU to look for obvious faults and found none. Everything is working fine now...Strange.
I had a hard disk taken out by a power spike. I also had firmware corruption that required reflashing the EEPROM, but I don't remember if that was related to the PSU. They draw 30 watts when switched 'off' so it's definitely worth turning them off when not in use.
Twitching fans might indicate over-current protection kicking in.
I know one thing however - switched mode PSU experts are few and far between!
Mark
--
http://www.wickensonline.co.ukhttp://hecnet.euhttp://declegacy.org.ukhttp://retrochallenge.nethttps://twitter.com/urbancamo
On Wed, 5 Feb 2014, Google wrote:
On 5 Feb 2014, at 05:26, Cory Smelosky <b4 at gewt.net> wrote:
So, my VAXstation 4000/60 shut off and I had difficulty getting it to turn back on. Seems like a not-to-abnormal situation, right? PSU's getting old or overheating protection kicked in? That's not the weird part!
It may have overheated, I do know power switches can get flakey in VS4000s but the symptoms tend to be it won't turn OFF not on't turn ON but I guess if the switch is not contacting correctly it could be that too. the switch isn't, as far as I can tell, directly controlling the current, it is in some kind of latch circuit that unlatches when you flick it to 'OFF'. I think it's an early 'soft-power' implementation.
I moved the UPS further away. Very possible it got bumped and got too close to the UPS.
Full explanation:
http://dectec.info/vaxstation-4000-power-switch-issues-and-cleaning/
Thanks!
Also talk to Mark Wickens, he's had a few issues with 4000/90 PSUs.
When I rebooted the system...it had decided it was suddenly /2015/ and all the licenses expired. I've heard of systems resetting to the past when something happens...but NEVER the future!
That is odd, but Mark W remarked on my blog about one PSU failing and spiking his 4000/90 so badly it fried several parts, it's possible it could have upskittled the TOY clock?
Fans would click for a second and then kick right back off. It stopped doing that until I removed all the drives. Plugged 'em back in after opening the PSU to look for obvious faults and found none. Everything is working fine now...Strange.
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On 5 Feb 2014, at 05:26, Cory Smelosky <b4 at gewt.net> wrote:
So, my VAXstation 4000/60 shut off and I had difficulty getting it to turn back on. Seems like a not-to-abnormal situation, right? PSU's getting old or overheating protection kicked in? That's not the weird part!
It may have overheated, I do know power switches can get flakey in VS4000s but the symptoms tend to be it won't turn OFF not on't turn ON but I guess if the switch is not contacting correctly it could be that too. the switch isn't, as far as I can tell, directly controlling the current, it is in some kind of latch circuit that unlatches when you flick it to 'OFF'. I think it's an early 'soft-power' implementation.
Full explanation:
http://dectec.info/vaxstation-4000-power-switch-issues-and-cleaning/
Also talk to Mark Wickens, he's had a few issues with 4000/90 PSUs.
When I rebooted the system...it had decided it was suddenly /2015/ and all the licenses expired. I've heard of systems resetting to the past when something happens...but NEVER the future!
That is odd, but Mark W remarked on my blog about one PSU failing and spiking his 4000/90 so badly it fried several parts, it's possible it could have upskittled the TOY clock?
--
Mark Benson
http://DECtec.info
Twitter: @DECtecInfo
HECnet: STAR69::MARK
Online Resource & Mailing List for DEC Enthusiasts.
So, my VAXstation 4000/60 shut off and I had difficulty getting it to turn back on. Seems like a not-to-abnormal situation, right? PSU's getting old or overheating protection kicked in? That's not the weird part!
When I rebooted the system...it had decided it was suddenly /2015/ and all the licenses expired. I've heard of systems resetting to the past when something happens...but NEVER the future!
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
I was able to get outgoing mail working without TOPS-20 being the
wiser. I'm now running "redir" on my local emulator host (linux) to
listen on port 25 and send all connections to my external mail server
(a site out on the Internet) on port 2525 (I set postfix to listen on
port 2525 in master.cf, in addition to the default). Then, on my
local firewall, I'm using iptables to send ALL outgoing connections to
any host at port 25 as a PREROUTING NAT rule to be diverted to my
emulator host's port 25, to hit the "redir" process. In this way
TOPS-20 can send mail to any valid e-mail address and have it
transparently diverted to my external mail server's port 2525, where,
subject to postfix's approval, it will deliver the mail. It just took
some ingenuity to work around Comcast's silliness.
-Mark
On Mon, Feb 3, 2014 at 11:08 PM, Cory Smelosky <b4 at gewt.net> wrote:
On Mon, 3 Feb 2014, Mark Abene wrote:
Here's the path madeline.gimme-sympathy.org takes: (4.3BSD running qmail. MX
record exists! Don't email b4@ yet...I haven't figured out incoming emails
yet):
madeline (10.10.3.2, qmail) -> frontgate/mercia (10.10.0.2, postfix) ->
mailer.gewt.net(external, authenticated relay, also postfix) -> THE
INTERNET.
The return path is MUCH simpler.
Where's the mailer config on TOPS-20 located?
-- Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On Mon, 3 Feb 2014, Mark Abene wrote:
Here's the path madeline.gimme-sympathy.org takes: (4.3BSD running qmail. MX record exists! Don't email b4@ yet...I haven't figured out incoming emails yet):
madeline (10.10.3.2, qmail) -> frontgate/mercia (10.10.0.2, postfix) -> mailer.gewt.net(external, authenticated relay, also postfix) -> THE INTERNET.
The return path is MUCH simpler.
Where's the mailer config on TOPS-20 located?
-- Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
my two choices for Internet in San Jose are Comcast and AT&T Uverse.
If you can get DSL service (and if you can get Uverse then you can,
right?) then you can sign up with sonic
www.sonic.net
That's what I use, and I live in Milpitas. Best ISP ever...
I only wish I could get Fusion service here - then I could dump AT&T
completely.
My suggestion about mail forwarding was completely off base. Sorry - I
must have been sleepy.
Bob
On Sun, Feb 2, 2014 at 7:42 PM, Bob Armstrong <bob at jfcl.com> wrote:
Comcast is evil. Evil.... My first suggestion would be to get a real ISP
:-)
Unfortunately my two choices for Internet in San Jose are Comcast and
AT&T Uverse. And being that Uverse is a totally unknown quantity, I'm
hesitant to gamble with switching. :)
Having said that, how about setting up a Linux box as a relay on your own
internal network? It can accept incoming SSL mail on, say, port 465 and
then just forward it over your local network to TOPS20 on port 25. Or does
COMCAST block port 465 too?
Incoming is no longer an issue, just outgoing. For the incoming, I
created a postfix transport on an external mail server that I control
out on the net, which sends incoming mail to my DDNS address on an
alternate (unblocked) port, which my firewall forwards to port 25 on
TOPS-20 and all is well. Outgoing mail is another story, since
Comcast blocks outgoing to any destination port 25. This is a little
trickier, hence why I asked about the mail relays internal to HECnet.
In the meantime, I'll see if I can create an outgoing rule in my
firewall to send all outgoing TOPS-20 mail to the same external mail
server, which I can have listen on another port in addition to 25 and
serve as a smarthost.
-Mark
Bob
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On Behalf
Of Mark Abene
Sent: Sunday, February 02, 2014 7:12 PM
To: hecnet at Update.UU.SE
Subject: [HECnet] TOPS-20 outgoing mail relay config
Long story short: after moving to the other side of the country a year ago,
I finally (recently) had time to get my TOPS-20 system (KLH10) back up and
on HECnet. Only to discover that Comcast (cablemodem service), in their
wisdom, blocks port 25 in and out, in an effort to "combat spammers".
This rather complicates my e-mail setup in TOPS-20. For incoming, I simply
created a transport relay on an external mail server that I control in order
to deliver mail on an alternate port number that isn't blocked, and then
simply port forward it from my firewall to tcp port 25 on the TOPS-20
server. That was the easy part. Outgoing mail from TOPS-20 is proving to
be a bit more difficult. I was thinking I may be able to intercept all
outgoing traffic from TOPS-20 bound for port 25 at whatever smtp server, and
redirect it to a single smart host (mail server on the net which I control),
on an alternate port that isn't blocked. Might work.
The reason I'm asking about this here, is I'm curious given the mail
transports at LEGATO and CHIMPY on HECnet, if anyone has tried simply
setting one of them up as a "smarthost" relay for all outgoing mail in
TOPS-20, perhaps via the <mail>mailer-relay-info.txt config file.
Unfortunately the syntax of mailer-relay-info.txt is a little confusing;
despite the explanation in the file's comments, there are no clear examples,
and trying to follow the logic of the options in the MMAILR MACRO assembly
code is less than helpful in modern mail server context (transmogrify??
who's domain??).
Any suggestions welcome.
-Mark
