PDA

View Full Version : H7864 power supply schematic [BA23 Box]



RSX11M+
April 15th, 2011, 11:55 PM
I'm hoping there is an electronic documentation set online that I haven't found for the H7864 power supply.

My BA23 box just blew one extravaganzaic fashion], and I need to play Mr. Fixit.

I have a couple of these boxes, so I probably should have tried to get these before, but I never had one go kaput on me. From the color and appearance of the smoke it produced, I expect it's a bad electrolytic capacitor. If it didn't take anything along with it, I may not need a print set to fix it.

My philosophy is that if I get the documentation first... I won't need it.

Does anyone have any suggestions?

Lou - N2MIY
April 16th, 2011, 03:50 AM
I have scoured the internet looking for that print set. If you look closely, that power supply was made by Astec for dec. You probably already see AA12130 as the Astec part number. If you can find the print set it would be very helpful to many people. I have an untested "spare" that is probably broke. 9 times out of 10 with switching supplies, I've found it's the main switching power mosfet to blame.

Lou

RSX11M+
April 16th, 2011, 06:44 AM
As always, thanks Lou.

I think I'll look in the Fiche set next. I just wanted to be sure I wasn't resorting to that unnecessarily.

Astec eh? No I didn't see that... I haven't taken it out yet. It surprised me so much when it popped in a "power-off" state, and I didn't feel like diggin into it right then. Had to ventilate the whole house overnight - busy with that.

Lou - N2MIY
April 16th, 2011, 09:46 AM
Well, if you haven't taken it out of the chassis, then you may find that it's not the power supply, but the wiring "harness" from the power supply to the backplane.

There is a known problem in early BA23s where the +5V connector overheats at the power supply end. I had this problem in one of mine.

Google "BA23 fire hazard" and you will come to this: http://www.people.vcu.edu/~agnew/MVAX/9412_DECUSERVE_JNL.HTML

Read it a little bit to learn all about the problem. I had it happen to me!

Lou

Lou - N2MIY
April 16th, 2011, 09:55 AM
However, yours blew up when the power was off. So, it's probably not the "BA23 Fire Hazard" problem.

I have also had power supplies blow up when not switched on. It has always then been the line filter. It's still energized on many dec power supplies, even when the power switch is turned off. I had this happen on two decmate power supplies. I just looked in the end of my "spare" H7864, and sure enough there is a little line filter board behind the IEC power socket. Although smelly, it should be easy to repair.

Lou

RSX11M+
April 16th, 2011, 10:12 AM
Yeah... I had a similar thought, the line filter fits pretty well. It didn't even pop the breaker when it happened. Time will tell. I'll pull it tonight.

Cheery thought.

I don't think I ever heard about the BA23 fire hazard thing before. I'll read up on it - Thanks.

RSX11M+
April 16th, 2011, 05:43 PM
Ok... A few updates.



Having read the posts at the link you provided, I can say that only one of these "Fire" hazzard items is familiar... that of the AC line chord being too light. Fortunately, this is something that was recognized back in the day.



The others, about cables and connectors between the backplane and the supply being too light, is news. Looking at mine, in the now disassembled BA23, these appear to be sufficient. I will check further in this and my other box, just to be safe.


I have now removed the H7864 supply [ASTEC stamped REV-A]. It is a very nice clamshell type metal enclosed design, which has 3 major board assemblies.



AC Line filter Board
High Voltage Rectifier / Filter Board
Main Low voltage switching supply



It is rated as a "250Watt" supply which outputs 36A of +5v, 7A of +12 [264 Watts already], and has additional lower current 5V and 12v rails. It also provides 12VDC power [0.5A each] to the 2 main system - 4.5" diameter "MUFFIN" fans . Equipment catalogs rate these at 110CFM, which is small by today's standards for the amount of noise they make.

Input ratings for the supply / box are labeled as 345 watts. This is a little confusing in that it appears to be a 110/220 Volt design with the typical embedded slide-switch voltage selector but the same 7A / 10-second series breaker (?).
With a 7A breaker this amounts to 770Watts @110v and 1540@220v. I now know why the breaker [U]never acts on these before major damage occurs.
It differs from modern [ATX] supplies primarily in that it also provides control signals: DCOK [DC OK], POK [Power OK] and a signal pulse for the LTC [Line Time Clock]. It also has a rear panel connector for "Power Sequencing" with other supplies, to deal with "inrush" issues in large racks.

The "Main Low voltage switching supply" board assembly is comprised of 4 more minor PC boards, mounted to a 5th main board, which has all the DC output connectors and cables on it, along with 3 transformers and other components.
My initial survey of the unit indicates the problem does indeed appear to have originated in the AC Line filter board, and is evidenced by an explosively destroyed 0.47uf non-polar capacitor. This board has several other components, among them - MOVs which I know from experience do not age well. It may be that one of these MOVs shorted in failure, and put 110VAC directly across the capacitor.

I will need to disassemble and analyze the circuit a little to determine the truth.

So far, it appears this is the best possible failure that could make smoke. More updates and photos later.

RSX11M+
April 17th, 2011, 11:10 PM
Photos -

Supply as removed:

http://96.11.235.30:9080/supplemental/H7864/H7864_a.jpg


Opened, with Line Filter Board area indicated:

http://96.11.235.30:9080/supplemental/H7864/H7864_b1.jpg


Line Filter close-up with bad capacitor C102 0.22uf [note C101 appears to be the same type but 0.47uf value]

http://96.11.235.30:9080/supplemental/H7864/H7864_line_filter.bmp


I am trying to decide between repairing the line filter board [replace both caps] and replacing the entire Line Filter with a CORCOM encapsulated unit.

I think it will be the repair, and inclusion of a smaller value, more appropriate breaker. I'm thinking 3A / 10 sec.

RetroHacker_
April 18th, 2011, 04:28 AM
Those clear poly caps fail like that with age. Just replace the caps and you should be fine. This is a very common failure in Apple II and Apple /// power supplies. See how the .47uf one is starting to crack? It's about to blow too. Replace it while you're at it. You want to use line rated X2 capacitors.

-Ian

RSX11M+
April 18th, 2011, 11:36 AM
RetroHacker_ Thanks for weighing in.

I could use a little guidance...


C101 appears to be Line-to-Line... so X2 would be fine there.
C102 [blown] is Line to Ground, so should this be Y2?

I found another of these types on the High Voltage Supply board too, and will look at it next, before ordering parts.

So far I have these candidates from Digi-Key


MKP 339 X2 0.47uf (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=BC2590-ND)
MKP 338 6 Y2 0.22uf (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=BC2605-ND)


Thanks very much.

RetroHacker_
April 18th, 2011, 11:51 AM
RetroHacker_ Thanks for weighing in.

C101 appears to be Line-to-Line... so X2 would be fine there.
C102 [blown] is Line to Ground, so should this be Y2?



That would be correct. Truth be told, I've used "whatever was handy" in the past - it works fine. The primary difference between the two is that Y2 caps are specifically designed not to fail shorted. But your research is correct - X2 caps should be used line to line, Y2 line to ground. You can also use Y2 caps in line to line circuits - they're just bigger and more expensive usually.

I tend to salvage these kinds of safety caps from more modern computer power supplies - the new ones are a heck of a lot nicer than those older clear ones. I've seen many of the clear block caps blow up, but never seen a newer one fail.

Lovely smoke they make, isn't it? Takes a while to get the smell out of the room, that's for sure.

-Ian

RSX11M+
April 18th, 2011, 11:58 AM
ROOM ??? - LOL

Sucker stunk up the whole house, and woke the old lady out of a dead sleep!!

Man-o-man did I pay for that!

Ok, thanks for the support... I'll keep ya posted. Maybe I'll finally cannibalize some old junk I have lying around for the parts. Heck, there's enough of it. :rolleyes:

Comment: I'm laughing at myself here referring to the "New Old Junk" as junk to be cannibalized to repair the "Older Precious" non-junk. Really tells me where I feel we're at today, eh?

RetroHacker_
April 18th, 2011, 12:09 PM
ROOM ??? - LOL

Sucker stunk up the whole house, and woke the old lady out of a dead sleep!!

Man-o-man did I pay for that!

Hahaha. Reminds me of when I was testing out a new find - a Bell & Howell black Apple II+. I was eager to play with it, and I set it up on the kitchen counter (workbench was full, after all). Machine works for about twenty minutes, then *PFFT!* *FIZZ!* and the machine is engulfed in smoke...

I was politely reminded that the kitchen counter is not a workbench...

But yeah - I've pulled suitable replacement caps from junk PC power supplies. They use the same sort of safety caps, and there aren't too many different values.

-Ian

RSX11M+
April 18th, 2011, 12:14 PM
Hahaha. Reminds me of when I was testing out a new find - a Bell & Howell black Apple II+. I was eager to play with it, and I set it up on the kitchen counter (workbench was full, after all). Machine works for about twenty minutes, then *PFFT!* *FIZZ!* and the machine is engulfed in smoke...

I was politely reminded that the kitchen counter is not a workbench...

You, my friend, have EXTRA LARGE "Attachments"!

RSX11M+
April 18th, 2011, 08:47 PM
Found one more of these types in the HV supply, across the bridge rectifier on the AC side.

Interesting note: The two caps that are "ok" are marked "X2"... the one that failed... not marked "X" at all. Same encapsulation and P/N series... just no "X" rating.

I just looked at one of my failed PC ATX supplies [stuck fan] to cannibalize... no input suppression or line filter whatsoever. Direct AC line to the input bridge rectifier. Absolute junko. "Disposable society" design mentality.

jackrubin
April 30th, 2011, 06:38 AM
RSX11M+,

Since I've got a couple of these systems as well as a "standby" power supply, I'm interested in proactively replacing the line filters (since I've also been known to work in the kitchen and my "attachments" are only normal size). You've listed C101 and C102 but also mentioned a third cap in the HV supply. Can you provide a final tally of caps by schematic reference and value?

Thanks,
Jack

RSX11M+
May 2nd, 2011, 12:39 AM
Here you go Jack -

The additional capacitor on the High Voltage rectifier board is C201, 0.1uf and appears to be marked X2 rated.

It can be seen in this photo on the left end of the board along the bottom of the photo.

http://96.11.235.30:9080/supplemental/H7864/H7864_b1.jpg


It's the one immediately adjacent to the black connector with the blue wires.

Sorry it took a couple days to get back to this... been a busy week what with the weather and so on.

jackrubin
May 2nd, 2011, 03:43 AM
Thanks! Glad things are OK with you. Chicago has been uncomfortably wet but no big problems here.

RSX11M+
May 2nd, 2011, 02:14 PM
As I've previously posted (http://www.vintage-computer.com/vcforum/showthread.php?24876-H7864-power-supply-schematic-BA23-Box&p=175857#post175857) in this thread, I cannot rationalize DEC's [or ASTEC's] choice of a 7 Ampere AC line breaker in this power supply.

Even at 110VAC, the current it permits is absolutely ridiculous considering the supply's maximum ratings, to say nothing of the 220VAC case.

Here is a summary of those ratings:
242 Watts MAXIMUM Output Power with Forced Air Cooling [sticker]


180W - 5V@36A
084W - 12V@7A
012W - 2x12V@0.5A [fans]


276 Watts total
There is another sticker on the supply which I take to be the Input Power rating

This states:


345W - 4.4A @ 120VAC or 2.2A @ 240VAC



Note: I cannot see how the breaker would ever act in a 240VAC situation. This design is just plain wrong by everything I know.

Re-Engineering Discussion:
However, 4.4A @ 120VAC = 528 Watts <<-- Way too much

Even if I use the +/- 20% rule for line voltage [this is how I was trained to design] 4.4A is 440 Watts @100VAC - still substantially over current. 3.5 Amps would be more like it.

At nominal 120VAC 3.5A would still be 420 Watts. <-- Still too much

So, I am soliciting advice. I will be making the same alterations to all my H7864 supplies.

My first inclination is go with a 3A breaker on the AC input, and to add a "fast blow" fuse on the DC output of the High Voltage supply. [I'll size that fuse based on component ratings - probably the bridge rectifier] I will also be disabling the 120VAC/240VAC switch and dedicating the supply to 120VAC.



3A @ 110VAC [my nominal line voltage] is 330 Watts.


The following is the manufacturer's guidance for the W58 Series Tyco breaker I'm considering.
http://96.11.235.30:9080/supplemental/H7864/W58_data_tyco.jpg

However, after seeing this data, I'm wondering if I should go still lower. Would that be too crazy?

Opinions?? ...and please, Be verbose - give your reasons, even if they're "religious" ones. ;) You won't hurt my feelings.

Lou - N2MIY
May 2nd, 2011, 05:08 PM
It seems worth putting your amp clamp on the line cord and capturing the power-on current transient on a storage scope. Then you would have an idea of the real system I-T curve that you can use to guide your breaker selection.

I have thought about doing this myself. I have a heavily loaded BA23 in the garage with an 11/53, two memories, RLV12, KDA50, DELQA and more in there. I don't have a storage scope though (and I am not buying film for the scope camera!)

Lou

RSX11M+
May 2nd, 2011, 07:43 PM
Thanks Lou.

I plan to do a few things...


Load up a chassis [probably using a dummy load - no sense risking real hardware] and get some actual efficiency numbers / curves.



I have a clamp-on AC ammeter - no sweat. Storage scope - good idea. The timing of the breakers is pretty common - 200% over current for at least 10 seconds. That feels like plenty to me, to get beyond any power-on transients.



I actually want to test their breaker, and my replacement if needed. It's hard to believe the numbers I'm reading. ** No longer needed - See below


That's the kind of thinking I'm looking for.



More supporting data: ...The original part is indeed a 7A breaker. [I think this will get you a datasheet (http://airpax.sensata.com/site/utilities/eliterature/pdfs/snapak.pdf)]



AirPAX PR11-1-7.00A-XX (http://ww3.airpax.net/CC_host/pages/custom/templates/airpax2/configure.cfm?cc_nvl=%28%28cpc,PR11-1-7.00A-XX-V%29,%28CC,production,P,snapak_pr%29%29)

The Manufacturer is still in business. [probably not in Maryland any more though] However, all other factors being equal - it does appear to be a very nice 7 AMPERE breaker.



This means at 120VAC the power supply would be taking 1134 Watts [+328%] to trip instantaneously at turn on, and need to go over 840 Watts to trip at all. That's 240% overcurrent!!



Even worse, at 240VAC those numbers are 2268 Watts / 1680 Watts ZOWEE - I'm rolling on the floor! [ 925%/500% !!]]


I cannot believe this of DEC. Anyone else have this reaction??

RetroHacker_
May 3rd, 2011, 04:48 AM
I cannot believe this of DEC. Anyone else have this reaction??

It does seem a bit high... but there may have been a reason for the madness - like to avoid tripping in brownout conditions, or during startup with a loaded system, heavy hard disk seeking, disk spinup, etc. Or, they could have just had a whole bunch of 7 amp breakers laying around and used those.

But truth be told, the purpose of the breaker on a switchmode power supply is to pop in case of catastrophic failure. If you short the output terminals, the switcher is supposed to just shut down - not overload. The only thing that would pop the breaker is a failure of the main rectifier or something... and that would draw a hell of a lot more than 7 amps.

And even the failure of those yellow caps doesn't usually draw enough current to blow the fuse on other, smaller supplies. The Apple II power supplies have a small glass fuse (forget the value, 1 amp? 2amp? It's not a lot) that doesn't usually go when those caps burn.

Personally, I'd trust DEC/Astec to know what they were doing, but then again, I'm not as well versed in electronic theory as you are. I just know what I've seen from years of fixing things, and I've seen some really, really stupid things (just look at the schematic for the original Wiliams Defender video game power supply... I dare you. No overvoltage protection at all), I think this one is a fairly minor oversight.

-Ian

Matlock
September 23rd, 2016, 05:50 AM
RetroHacker_ Thanks for weighing in.

I could use a little guidance...


C101 appears to be Line-to-Line... so X2 would be fine there.
C102 [blown] is Line to Ground, so should this be Y2?

I found another of these types on the High Voltage Supply board too, and will look at it next, before ordering parts.

So far I have these candidates from Digi-Key


MKP 339 X2 0.47uf (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=BC2590-ND)
MKP 338 6 Y2 0.22uf (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=BC2605-ND)


Thanks very much.

RSX11M+,
I bought a spare BA23 box at the VCF Midwest a couple weeks ago and started to test it out last night. I put a 1 amp draw auto light bulb (current draw verified at 5V with lab power supply) on the 5 Volt output of the BA23. I also put a 0.3 amp load (small PC fan) on the 12V side. The BA23 started fine and I measured 5.05 Volts at the bus. The fans were both running fine and I decided to let it run for a few minutes.

After about 5 minutes, the C102 capacitor exploded (rather spectacularly) just as was shown in your photos.

Were you able to repair your H7864 power supply by replacing the C102 & C104 capacitors? Where did you find suitable replacements and what were their part numbers?

Thanks for your help,
Matlock

jackrubin
September 23rd, 2016, 05:59 AM
Nice!

You not only got a good system at a great price, you got extra added excitement for no additional cost!
VCF-MW rocks!

:D

Come back next year, ya'll!

Jack
Proud MidWesterner!

Matlock
September 23rd, 2016, 06:44 AM
Nice!

You not only got a good system at a great price, you got extra added excitement for no additional cost!
VCF-MW rocks!

:D

Come back next year, ya'll!

Jack
Proud MidWesterner!

Jack,
The 11/23+ & 256KB memory boards that were in this BA23 box worked fine (I tested it in my BA123 box) so that was great. I just got around to testing the new BA23 itself and it did create some excitement. It sounded like someone put a few fire crackers in the PSU. The smoke and a few pieces of blown capacitor shot out through the fans and I shut it down as quick as I could.

Best,
Mark

m_thompson
September 23rd, 2016, 08:08 AM
The notoriously unreliable Rifa X and Y caps are for EMI suppression. If you don't care about noise from the power supply going into the power distribution in your house, you can remove them and leave the, out.

wa2flq
September 23rd, 2016, 08:47 AM
The notoriously unreliable Rifa X and Y caps are for EMI suppression. If you don't care about noise from the power supply going into the power distribution in your house, you can remove them and leave the, out.


That settles it.. Time to schedule a preventative maintenance on mine.

Is there a list of common DEC power supplies that have these troublesome components?


Jerry

Matlock
September 23rd, 2016, 08:54 AM
The notoriously unreliable Rifa X and Y caps are for EMI suppression. If you don't care about noise from the power supply going into the power distribution in your house, you can remove them and leave the, out.

Thank you very much for this information. This will allow me to at least verify that there are no other problems with the power supply before doing anything else.

In general, I wouldn't care about some noise on my home's AC unless it would interfere with the ethernet over AC modules that I'm using to get internet access into the basement. I guess pull those two capacitors and see what happens, if it causes a problem I can eventual find good replacements and if it doesn't cause a problem, even better.

Thanks again!
Mark

Matlock
September 26th, 2016, 09:39 AM
The notoriously unreliable Rifa X and Y caps are for EMI suppression. If you don't care about noise from the power supply going into the power distribution in your house, you can remove them and leave the, out.

This weekend, I pulled the Rifa X and Y capacitors and the power supply is running fine. It did not seem to cause any issues with the ethernet over AC units I'm using, so I plan to leave it this way at least until I find a reason to replace those EMI suppression capacitors.

Thanks for the help!

Matlock

m_thompson
September 26th, 2016, 10:15 AM
This weekend, I pulled the Rifa X and Y capacitors and the power supply is running fine.

The FCC is going to hunt you down now. ;-)