The Big Score! Mountain Expansion Chassis!

Ok.

One other super simple test you can do is a continuity test for each of the wires in the 16 & 50 conductor ribbon cables.

Tedious. Sure. But in troubleshooting it is best cover all bases. As we progress through the job it gets easier. Parts and sections that are working are eliminated from the suspect list.

Set your DMM to the diode or continuity setting. Probably the one that looks like a rocketship with radar emissions. Touch the probes together and listen for a beep or significant change in the display.

For each pin or hole in one side of the cable there should be a beep when you find it on the other end. Just go through them one by one.

Keatah,

I replaced 2 of the center plastic standoff's with some 1/2 inch metal ones that I ordered from Jameco. I'd rather have the plastic ones, but I didn't have any in my stash.

Yes. The existing ones are getting a bit stiff with age, but I managed to reuse the six (3 each side) original ones. :)

Hello Steven,
please excuse my intervention in this point....
if you replaced the plastic standoffs by metal ones - did you take care that the srews are isolated from the board and the chassis ? Otherwise there might be a risk of a kind of shortcut between the board and the chassis....
large areas of the layer of the board are connected with ground - but probably not in every place ....
sincerely speedyG

@Speedy.

Yes I immediately thought about that too. The only electrical difference is that the chassis motherboard is now making the metal case into a ground. In our industrial/laser environment back then it was an issue! It would act as an antenna and pick up electrical noise from the shop. And we had all sorts of noise going on. So we had to use plastic standoffs and ground the case "manually".

Here in the home environment I don't think it will be a problem. All the holes where the metal standoffs go into are one and the same ground. It's just that the metal box is now at ground potential.

It is useful to note that the power brick is AC output and allows the Apple II+ power sub-system to set the ground. The power supply in the chassis is floating so to speak. Ground pathways in the chassis all come from the Apple II itself.

It would not do to have the power brick provide ground. Too much potential (no pun intended) to make ground loops! And I hate ground loops.

I also looked at the misc fittings and other areas that could short against the metal box. I didn't see any.

@Steven.

Troubleshooting outside the box is a good thing. I may ask you to make measurements on the backside. I'm assuming these are the standard hex standoffs like in a PC, just longer. I'm also assuming no traces near the standoff areas are damaged.

I'm also going to ask you to make Two more preliminary scouting measurements. The first of these 2 additional measurements is because you drilled out the holes when you replaced the 2 capacitors. Reference messages #39-43. It sounds like you just drilled enough to widen the existing solder, and that there is/was still solder left over. Thus the drill bit would not have affected any vias (feed throughs).

But let's be thorough and do it right.

I was going to mention it earlier, but it would have gotten lost in the busywork of gathering information and building the schematics.

Now. Now that you've got the right equipment (DMM) we can get rolling. The negative lead of each capacitor (C1 & C2) connects to ground via 4 little traces (2 per cap). These are somewhat visible in message #54-55 picture T-bcd1234_1.jpg. The negative leads also connect to DB1, the black square thing. This is visible in message #54-55 picture Topviewd2.jpg. That black thing square thing is a bridge rectifier. The capacitors and this bridge rectifier work together in concert to transform the AC power into DC power that everything else can use.

Well, these ground connections are under the capacitors. And you probably can't stick the probe underneath. So we'll do it another way. Flip the board over, and where you marked "-" touch a probe tip there. See message #54-55 and picture BottomsideTotalSM2.jpg. Yep. Just set the DMM to continuity (rocketship and radar on my Fluke meter). Red to the "-" marked trace, and black to a ground point such as the mounting hole or TP that is marked ground. You should get a beep or 0.00 reading.

I will also note that there is a backup trace that connects this part of the circuit. It is under DB1. Pin #4. IIRC. So in the event the traces are damaged from the drilling, and your board has this backup trace, you'll still be fine and we can move along.

The second continuity measurement is for the fuse. Make sure its good. I ask this because in message #130 you had the chassis powered and working. Then it was dead soon enough. So just touch one lead to each end and listen for the beep. That indicated a short which is good in this case. Perhaps a visual inspection will do here.

And here is a cautionary note. You will want to be sure the capacitors are discharged when making continuity measurements. And of course all continuity and resistance measurements will be carried out with the power off! And chassis unplugged from AC and disconnected from Apple II+. You may want to build a capacitor discharge tool like so --
http://www.ifixit.com/Guide/Constructing+a+Capacitor+Discharge+Tool/2177/1
-- Any other method will likely shoot solder blobs across the room. No kidding. I've seen it happen. And the resulting static blast can and does cause latent damage in CMOS. And in TTL to a lesser extent.

I personally recommend using a 1K or 2K ohm resistor, touch it across the capacitors and they will discharge with no sparks or smoke. Takes about 30 seconds. And you can watch progress with your DMM. Once it drops down to 1 or 2 volts, that's one or two, then you can safely short them with a coin. You'll get microscopic sparks for an instant. But nothing bad. The capacitors seem to like retaining a residual charge. So a solid short for a few minutes is a good thing. And this activity corresponds to this circuit only. Bigger and badder capacitors need different resistor values and methods.

You can see how much power is in them by measuring the voltage across + and - as seen in message #54-55 BottomsideTotalSM2.jpg. Measure them just like you would a battery.

A functioning chassis seems to discharge on its own in a minute or two.

I've been able to measure almost 40volts potential difference at various parts of the supply circuit. Maybe there could be higher potential elsewhere. So a note of caution is in order. This is a power supply!

If you're not familiar with electrical safety I would suggest you get a book and read up or attend a class.

Some of my good practices include safety glasses and 1 hand in the pocket and working on wood or plastic surfaces.

http://electriciantraining.tpub.com/14175/css/14175_58.htm
http://www.allaboutcircuits.com/vol_1/chpt_3/9.html

I'm not responsible for any damages or mistakes or failure to follow proper safety procedures resulting in injury or property loss. Sounds ominous, but in today's overly litigious world things like this need to be said.

Now that I've scared the bejesus out of you let's get going and see what we can see! It's gonna be fun!

Print out messages #193, #198, #202, and this one, #206, and keep as a reference. Conduct the tests after you've familiarized yourself and have confidence in using the tools at hand!

@Speedy again..

I checked on my chassis motherboard, and as I suspected, all the ground points are common to one plane. There don't appear to be any shorting between regions or internal ground loops being created.

Continue to feel free and interrupt anytime with any suggestions. You may come up with something I might take for granted.

Also, in reviewing message #130 I don't believe that is normal behavior for a chassis. The disk drive should not spin up when the button is pressed. You need to do it like on the actual Apple II, PR#6 for example. And if the red chassis light is on, pressing reset on the Apple II should turn it off. You can re-enable it through a poke, or manually pressing the button again.

Pressing the chassis enable button should only make the red light on the chassis come on. Why the Duo-Disk started to run at that point I do not know yet. I tried my genuine bona-fide Duo-Disk just now. It worked like a regular drive and didn't behave like described in #130.

I loaded some software and nothing blew up.

@Speedy again..

I checked on my chassis motherboard, and as I suspected, all the ground points are common to one plane. There don't appear to be any shorting between regions or internal ground loops being created.

Continue to feel free and interrupt anytime with any suggestions. You may come up with something I might take for granted.

Also, in reviewing message #130 I don't believe that is normal behavior for a chassis. The disk drive should not spin up when the button is pressed. You need to do it like on the actual Apple II, PR#6 for example. And if the red chassis light is on, pressing reset on the Apple II should turn it off. You can re-enable it through a poke, or manually pressing the button again.

Pressing the chassis enable button should only make the red light on the chassis come on. Why the Duo-Disk started to run at that point I do not know yet. I tried my genuine bona-fide Duo-Disk just now. It worked like a regular drive and didn't behave like described in #130.

I loaded some software and nothing blew up.

Hello Keatah,
this is amazing.... up to my analysis in posting #185 i thought the engineers at Mountain Computer realized more logic
in the control.....
Should i correct that description of posting #185 ?
this seems to confirm my statement made above, that there is difference between theory and practival experience....
sincerely speedyG

You could update it.

Also, if you press and hold the button on the front panel and then turn on the Apple II+, it will boot from disks & controllers installed in the chassis.

There really is no "intelligence" in this box. It's pretty dumb.

We would enable the chassis via a poke to a memory location, the exact number is in the manual. Then we'd read or write our data to the optics control cards, then de-select the chassis.

Thus swapping which set of slots (internal or external) would be in use. I hope I can still find the old a/d d/a card plans and programming they did back then.

Pressing reset on the Apple II+ de-selects the box, and you must poke memory location or press the button to bring it back online.

@Steven:
Take your time and have fun with this.

Whenever I used my chassis in the past, I always used slots 1,2,4,5. And that was mostly dictated by software expecting other cards to be in 0,3,6,7. Memory, 80-Column, Disk, Clock, respectively.

The chassis will actually work in any slot. We can stick MHEC interface card in slot #1 of the II+. And we'll use a 48k configuration for testing and troubleshooting.

And you can put the controller card into any slot in the chassis except 0. We can place it in slot 6 for our testing purposes.

This will be my starting configuration for testing:
MHEC interface card in Slot #1 of the II+.
A disk controller card in Slot #6 of the chassis.
Duo-Disk or standard original-style Disk II will be fine.
No other cards in the system. A straight 48K system, no 16k cards. We keep it simple.

What make/model DMM you got? Maybe there are other features that can be useful.

And prior to beginning testing, you'll want to be sure your II+ starts up normally and can load stuff. A quick test of the Disk II drive and interface card here will add confidence to testing as we progress through things.

That will work.

Well Steven,
that sounds like that you are excellent prepared.....
anything else ?
OF COURSE ! GOOD LUCK and the silent prayer that the trouble is found and solved fast !
lol
sincerely speedyG

Nope. Looks like you've got everything ready to go.

The purpose of the measurements I'm asking for is to determine whether the chassis has a good power supply. That it powers up in conjunction with the Apple II+. And that it is within a reasonable spec. I don't recall Mountain ever specifying exactly what they are supposed to be. But I don't worry too much about that. As long as 5V is within a few tenths. And the 12V stuff is within a volt. Preferably erring on the greater side - so that as the chassis is loaded with cards it can bog down and sag its way into spec.

Am I correct in remembering you have already installed a set of new chips? Both in the chassis and the interface card? Would this be all the chips? Including the small 8-pin bugger near the front-panel connector?

And I assume all these have 74LS as part of the part number. Except for the 8-pin guy.

I also assume there are no bent pins or reversed chips.

Hello Steven and Keatah,

in fact the "8-pin guy" is a 555 Timer. It is manufactured even today from different companies, very simple and common ( like NE555 or CA555 or NS555 ).
you can find that guy in nearly every thing the determines a specific period of time within the range from several millisecondes to several minutes ( egg-cooking-timer, protoprocesstimer and so on ).

If Steven has a complete set of the chips used in the interface, most chips on the mainboard are availiable as spare,
because in the area close to the 50 pin flat cable they are repeating .... the only exception is the 74LS245 - a very common bus transciever the often has been used at interfacecards close to the slot connections and controlling the
data- or adressingbus - so maybe at least one aa a "interim spare" is availiable from another card....

and the 8 pin guy ? well if there is a old Apple II clone availiable - there are 2 of them:
one at the front right corner - location A13 and another at the front left corner at the location B3 !
So one of them might also be used as interim spare.... and the 74LS245 have been often used on different
old 80-col cards for the IIe and the IIe itself too ( close to the auxiliary slot).
Up to my experiance in most cases only one or 2 chips fail, because voltage walks for least resistance ( directing to shortcut or fail ) and therefor often even the neighbors of such chips are left without harm.

After the description i read my suspect is lead to one of the powersupply-branches .... favorably the positive 5 or 12 volt. If we know which one we might know which of the transistors or diodes has been gone "off duty" - depending to the outcome of the measurements. It might even be indicated by the guy getting very hot.... - but thats at the moment just a guess...

sincerely speedyG

Yep. There could be a chip that is running hot and maybe that could pull down a supply rail or cause it to blow.

Remember: This chassis did power up and the duo-disk light came on too. Ref msg #130. Supposedly with no activity on Steven's part. Which is odd, because, you need to hit the chassis select button or do a poke command to get it rolling. Turning on the computer with a controller card in the chassis should just start you in basic, no activity on the disk till you do something. You have to hold the select button in AS you turn on the II+ to make the chassis "present" or "give" the controller prom to the autostart rom to effect a boot IIRC. Either that or power up the II+ then select the chassis, the PR#6 to get the disk going.

But then it all got powered down and restarted. Then the chassis was seemingly dead. I assume the green light never came back on.

Whether the duo-disk made the recalibration clicks the first time I don't know. And Steven probably doesn't remember that detail. But the recalibration clicks are of course controlled by the 6502 controlling the stepper motor. 6502 just blindly executes enough steps to haul the head from yonder'track 34 back down to 0. And a few extra kicks for good measure.

Well.. we're starting fresh and will be armed with some measurements.

We will get some answers in the days ahead. And as always anyone can chime in with suggestions. It is summertime here and I'm spending less and less time with retrocomputing activities. A seasonal thing to be sure!

Very good. I am preparing 2 more diagrams for 2 more readings.

[b]EDIT ADD PICTURES IN MESSAGE![/b]

I would recommend doing a transistor test on Q7, the 2N2907.
I would also like to see if D9 is alright.

These are easy measurements. They are looking at the chassis' ability to sense when the
Apple II is turned on.

And they are to be done with your DMM in diode test mode (the rocketship icon). It's
located to the left of the hFE setting. Also, these readings should be done with no power
connected and the 50-pin cable removed from the chassis. If you want to know more read
about diode test in your DMM manual.

This is easy. And we're looking for a little over 0.6 indication in each of the 3
readings. See these diagrams.

Hello Sreven and Keatah,
the 0 Volts indicate that the 2 Power Mos FET´s are not switching.
so the mess is between input of the 24 Volt DC the rectifier and the switching transistors in the oscilator
below the 2 big capacitors exchanged.
I´ll upload the pictures and the further steps.

@Speedy:
Absolutely! I believe this to be a power supply issue. Not a logic problem. Not an
interface card problem. Probably not a cable problem either. I'm asking Steven to
passively check the power on sense circuit. Also note that it is at 25V, which means that
25V is going back to the Apple II (!!!) or there is an unbalanced ground in the chassis
supply.

+23V is present at the cases of Q1 and Q2 switchers regardless if the chassis supply is
engaged or not.

I may not be available for troubleshooting till Sunday afternoon. So you can carry the
flag.

@Steven:
At this time I suspect the power supply is out to pasture or it has a grounding problem.

I want you to verify that the ground of the chassis (a standoff mount hole) is
electrically connected to the Apple II power supply casing. Just make a quick continuity
measurement between the two. Do this with everything hooked up. But power off.

The Ground test should = 0 ohms (the horseshoe omega symbol) or a continuity beep.

You can also measure the resistance between the 5V TestPoint and ground. It reads about
380 - 500 ohms on my rig. This checks for shorts on the 5V supply out. It's good DMM
practice.

And if you could report back the prior three readings from msg #223 like so:
Reading #1
Reading #2
Reading #3

Lots of things. But we are making significant progress toward finding the bad part.

Hello to Steven and Keatah,
it took a while uploading because at the moment my hoster of the website is debugging on my behalf a database/PHP script, because i believe they transmitted to me wrong connectiondata... that blocks / slows down my access......
so now after the upload:

@Steven:
first the picture again with the demands from keatah:

Now rhe picture with your measurements:

AND NOW HERE WE GO WITH THE NExT STEPS:

take a measurement in AC Voltage from the powerbrick !



next step will be measurements from the SOLDERINGSIDE OF THE MAINBOARD because on top there are no good measurementpoints :

then the following step again back to the TOPSIDE ! BEWARE THAT AT THE TRANSISTORS EACH 2 (!) Measurements are required:

Explenation:
In fact at the 2 Transistors there should be a rectangle sqarewave present - but due to the fact that there is no scope availiable we have to make compromise....
MAKE AC voltage and DC Votage reading !
While making AC Voltage reading you should nearly read the Voltage Peak to Peak !
In DC Voltage the DMM will make an average middlevalue of that voltage dependent to the duty cycles and you will get far less of summarazied current/voltage out from "on duty cycle" ONLY. Therefor you should get some 45% to 55% from Peak to Peak reading ! ( If the Transistor in the oscillator are on swing )
If not the Transistors will remain in AC Voltage measurement nearly 0 Volt and in DC measurement tending to the Voltage they get at input !

Best will be report back by :
Picture 1: Measurement 1:

Picture 2: Measurement 1:
Picture 2: Measurement 2:

Picture 3: Measurement 1:
Picture 3: Measurement 2:

sincerely speedyG

Update by Feedback from Keatah and his measurements:

I followed along with the measurements you suggest in message #226. My readings were:

Picture 1 (power brick, no load)
28.63 VAC

Picture 2 (24v regulator and bridge out "-" ) (Apple II on and no cards in chassis)
Measurement #1 32.68 VDC
Measurement #2 23.75 VDC

Picture 3 (resistor) (Apple II on and no cards in chassis)
Measurement #1 23.75 VDC
Measurement #2 7.14 VDC

Picture 3 (power fets) (Apple II on and no cards in chassis)
Measurement #3 23.50 VDC and 19.52 VAC
Measurement #4 23.71 VDC and 19.10 VAC

With the chassis plugged in, and the Apple II off I get 23 VDC and 0 VAC equally on both power fets.

last line from his mail indicates that:
If Apple II is connected but Power is OFF at Apple II
- the Part of circuit with Transistors 2N2222A ( the 4 in the row below Voltage regulation IC with heatsink ) -
do not swing the oscilator and drive the big POWER MOS FET´s ( giving the values in brackets ).

Powerup at the Apple II and the oscilator should start swinging and the Power MOS FET´s should be
working on the coil - ( giving the values without bracket ) and then -
depending to the condition of the parts that follow up after the coil ( indicated by violett shading ) -
the values you measured in your first picture on top ( those with 0 Volt at testing points )
should change and become alive.
sincerely speedyG

Hello Steven,

that will be fine... just mail me the results....
Keatah mailed that he´s "off duty" till saturday or sunday.....
we can proceed anyhow....
sincerely speedyG

Good evening all,

@Keatah: I know what you mean about "Down Time". LOL
I have sent Speedy a picture of my DMM so he can identify the "Rocket Ship" that
Keatah is talking about.
I also asked him to post the picture with Rocket Ship location up here so that everyone can see it.

Steven :)

Well hello to all AFmembers,
after i got the mail i collected the manual from Keatah´s DMM for better explenation with picture.
Then i assembled the pictures from Keatah´s DMM and Steven´s DMM together and filled up with some
explenation. The following piccture should then explain everything well to the community.....
beware of the bottom note - it´s a very commen trap performed by a lot of user´s - therefor i
added that note.

sincerely speeedyG

I'm looking to do a diode conductance test in message #223.

Steven's meter has this function. At this setting, the meter puts out a small current and
measures if and when the diode conducts. When it does a small voltage is developed and
read. Typically 0.6 to 0.7 for silicon diodes.

So all you need to do is turn the meter to the green asterisk setting like so..

.. and probe the circuit like in the 2 pictures posted in message #223. 50-pin & AC
powerbrick should be disconnected when doing this test.

There will be 3 separate measurements here:
Reading #1
Reading #2
Reading #3

Here is a more in-depth explanation of a diode test: http://www.allaboutcircuits.com/vol_3/chpt_3/2.html

Hello to all AFmembers,
here the measurements taken by Steven from his box:
comments will be added from me below the pictures and Keatal will add his analysis in the next posting.

this shows that the powerbrick is O.K.
The output Voltage is within the expected range.

This indicates clearly the power regulation IC Q8 to be damaged and out of order.
At the input there should be the assumed voltage from the 2 big electrolytic capacitors....
normally it should be limited to a value below 40 Volts ( that´s the limit for the capacitors...)
and at the second point ( output of the IC there should be regulated 24 Volts DC )
There even is the danger that this was caused by the fact that at the capacitors the voltage
loaded up to more than 60 Volts and thereby exceeded the permitted limit of inputvoltage of the IC.
[u]There is Danger that the capacitors have also been damaged, because the limits of that capacitors
also have been exceeded beyond permitted limits ( limits is 40 Volts and actual there are 60 Volts.[/u]

therefor the voltages to be expected in this picture can´t become reality.... the IC does not issue
voltage high enough to get correct values.

but this picture at least confirms the diode to be intact.

also the value in this picture is within the expected area and confirms intact component.

* Due to the fact that the voltage might - under bad condition - load up at the capacitors to 60 Volts,
it might be recommended to take while replacing the capacitors instead of the 40Volt type the 63 Volt type.
This might ensure that the capacitors don´t get demaged inside at the isolation .....
100 Volt type should not be used otherwise the risk id to high that they might even load up to higher voltages
and then exceed the permited range for the Voltage regulation IC ( * ) even if a "High Voltage Typ" is used.
Up till tommorow i will check for such a High Voltage typ of regulator that could replace the used typ.
The used typ is by datasheet ( see in the MCEB.DOC - the link is in a earlier posting above )only permitted
to be used up to maximum of 50 Volts ! The "High Voltage typs are permitted to be used up to 60 Volts and they
have a "security margin" of at least 5% which would set the maximum limit up to 63 Volts. That would be just
slightly above the measured value.....

next update will be performed tommorow.....
sincerely speedyG

I agree on Q8 being faulty.

I wonder what's causing 60 volts to build up across the capacitors? Lack of load (faulty Q8)
or faulty bridge? Maybe both..

Hello to all AFmembers,
hello Keatah,

well i guess the bridge rectifier to be O.K. but that can be proved while purchasing the Voltage regulator and some fitting Capacitors.

I have been viewing the onlinepages at Jameco to check if they have high voltage typ of 24 Regulation IC in TO3 case.
I didn´t find them having one. Q3 should be replaced with a high voltage typ, because it seeems that if no load is in
box, it might under bad conditions happen, that the Caps load up to 60 Volt. The replacement typ has to deliver 1,5 Ampere. Anyhow viewing the Online pages i could not find any HGC of HGV typ of 24 Volt in TO3 case. Probably it should be a MIL-typ... ( Hughes ? Fairchild ? )

Keatah, can you add here a source for such a high voltage replacement IC in the USA ? It´s difficult to research such a source here from Germany.

The high Voltage of the capacitors is surely the result from no relevant amount of power pulled off from the box itself.

That causes the caps to load up that high.... under normal conditions i would guess 63 Volt typ of caps to load up that high... but if 40 volt typ can load up that high beyond their specified limit under specific conditions - i would surely misstrust the caps to be O.K. if only the IC has been replaced. At least that 60 Volt will be quite sure the suspect of killing the IC - because it´s beyond the limit of their specified input limit of maximum 50 Volt !

I just even guess the guys at Mountain Computer knew about this problem because they have choosen that 50 Volt limited typ - instead of the normal typ - that is limited at 40 Volt input.....

... under the given results from the measurements i guess it would be worth thinking and discussing if the cooling sink should be replaced with a version that is a little stronger / better with more "cooling fingers".
sincerely speedyG

I don't think we're going to easily find an original SiliconGeneral SG 340K-24 regulator.
The next best equivalent would be NTE1924.

Available here and perhaps elsewhere.
http://www.weisd.com/test/GenericParts_WEISD_view.php?editid1=SG340K-24
http://www.mouser.com/ProductDetail/NTE/NTE1924/?qs=%2fha2pyFaduiTBJpmXTxNTFk99rbhBEGH%252bsp0etfcxNs%3d

I'm not aware of any high-voltage version of the 340 series.

On Steven's board do the 2 large capacitors' negative leads and bridge out (-) connect to ground?

Hello Keatah,

i do know that there exist high voltage versions up to 60 Volt input voltage.... i have been working with that versions about 15 years ago.... Your proposal eqeals to the UA7824 and both are limited to 40 Volt input. I don´t believe that to be a good choice , if even the technicians from Mountain computer have decided to use a version limited at least to 50 Volt input voltage.....
The Versions i am talking about have been manufactured by the same guys that made the LM317HVK:

They also had a series of fixed voltage regulators ... at least 15 years ago they had....
and as far as i remember there were 2 companies that manufactured HV versions ( with up to 60 Volt input )
but i also know it was difficult to get them because they have been designed for MIL usage ( for example in tanks )....

Anyhow - if the replacment is to be carried out - one thing is for sure - we must hunt for a Regulator with permission of higher input voltage than 40 Volt!
By comapere to the original part it should at least permit 50 Volt input voltage !
sincerely speedyG

Back.. for a few days.. Now let's see what we can see..!

For parts sources you can try here -- http://www.eciaauthorized.com/

While the voltage regulator may be bad, I think we need to stop right here and ask
exactly why is there a 60 volt reading across those capacitors? How is that happening? From
practical experience I don't recall that ever exceeding 36 volts.

I also feel that substituting one of the LM117/317 adjustable regulators is going to become
too complex here for a number of reasons. The pinout is different from the SG340K-24 and the
LM117/317 are user adjustable whereas the SG is fixed at the factory. A reference circuit needs to
be designed so that the regulator knows where it's at. Vin and Vout are different in the pinout
diagrams. The LM117/317 has and adjust point in place of the Vin on the original SG.
is a ground.

Lm117/317 is not directly interchangeable with the original SG340K-24, whereas the NTE1924 is a
drop-in replacement. A reference circuit would need to be built. And then there is cost and
availability in small quantities to consider. For the full mil-spec part, expect to pay over $50,
and that's in a bulk order.

I still recommend the NTE1924. Mouser has it p/n 526-NTE1924 and it's $10.00 + s&h.

While it is not souped up spec wise. It is made with modern mfg processes and I suspect there's
more wiggle room than the datasheet says.

But we really need to figure out where the 60 volts across the capacitors is coming from.
As that is 2x over spec! Lucky they haven't exploded and bopped someone in the head!

Hello Keatah,
i have sent you a mail....

some part of the reply is due to misunderstanding of my posting...
i did nit make proposal to replace with 317 or 117...
they have been displayed to show, what i talk about, when i say "high voltage typ"......

the Q8 is definitly dead !

The 60 Volt result just from chatging up the "big caps" with no load,
due to the fact that the MOS FETs are switched off or
don´t work at the demanded level and don´t work "on the coil"....

In fact i also don´t trust the "big caps" anymore
due to the fact that they have been charged up far beyond the point
they have permission for ( they are speced only for 40 Volts ).

I´ll add some stuff related to the topic of the regulator tommorow too !
to avoid misunderstanding please also recognize the remarks made
in text and the explenations in the pictures too !
sincerely speedyG

@Speedy:
Sent you mail.

@Steven:
I think we will need to start isolating part of the circuit to figure out why the capacitors
are being pushed beyond spec to 60V.

This means learning to discharge capacitors safely. It also means taking the Q8 regulator
out of circuit. We are likely to recommend replacement of the regulator anyways. So it will need to come out.

I am going to use my junker chassis as a test rig and see if the caps stay at 34V without a regulator. I suspect they will.

Alright. With no load and no regulator in circuit the capacitors get pumped to 34 volts DC.
They stay there until something discharges them. They don't go any higher. I repeat-o, they
don't go to 60 vdc.

For completeness sake, the ripple is a negligible 7.8 mV.

I expect we are going to be removing the voltage regulator, testing the no-load voltage of
the capacitors, and replacing the voltage regulator with a new one.

I expect Speedy will chime in with the specifics.

If this fails to repair the power circuitry I would strongly like to continue
troubleshooting with a O'scope. While continuing tests can be made with DMM, it has to be
characterized to see how it responds to different frequencies and pulse widths and all that.
A lot of tedious work.

We also need to determine why the regulator gave up the ghost in the first place, if indeed
it is the regulator at fault.

Hello to AFmembers
and of course Hello Steven,
well first lets get the current task to be done:
Steven it is upmost important to remove ( unsolder and then remove screws to get the item away from board )
Q8 ( the SG340-24 ) with its cooling sink. Then after the IC has been removed
there should be the second measurement ( in the above posting the measurement at the soldering side )
should be repeated to find out if the Voltage rises up again to values above of 38 Volt DC.
To make the measurement reliable it is upmost important to:
1.unplug the MCEB
2. discharge the 2 "Big Caps"
3. connect the Ground of the MCEB to another reliable ground
4. power up the MCEB and repeat the measurements at the solderingside of the "Big Caps" ......
( same procedure like in description above )
and then immediatly after the values are collected to issue the result here.
This measurement will then lead to the decision if the "Big Caps" must be removed too and if some measurement
at the rectifier is required.

@ Steven:
i immediatly release this posting, so you can read it.
meanwhile i will edit by an Addin but that is not part of this step - it´s related to the search for a replacement of the damaged SG340-24. That part is very difficult to get and depending to the results of your measurements that are related to this task above - it will simplify or complicate the search for a replacement
of that regulator.

sincerely speedyG

UPDATE / ADDON:

@ Keatah,

i do assume that the Voltageregulator IC has been killed by the high voltage of 60 Volt...
that is far beyond the maximum inputvoltage and as explained i´m very sure that the decision for this chip
was not done by availiability but more to the fact that the Guys at Mountain Computer knew abot the problem / risk of higher voltages under bad condition.... it would have been much easier and far more cheaper to use a simple regulator like the in those days very common LM309 or similar typs. But they are all limited to only 40 Volt maximum input voltage.

The replacement for the IC will anyhow take several more days so that at the moment we can relax till we get the measurements from Steven. In the meantime just to spot it out:
there are 2 ways to keep away that dangerous "high voltage" loads from the following electronics behind:

1.option: replace the powerbrick from Mountaincomputer ( seems to be in reality 28 Volt /1,5 Ampere AC )
by a transformator that really only delivers 24 AC with 1,5 Ampere

2.option by adding a resistor with "high powerabsorbance" as i explained in my mail to you
Value should be something about 3,3 kΩ to 4,7 kΩ and at least 5 Watt - better 11 Watt .

In that case i would make a precise drawing how to perform this solution without any changes to the
mainboard. But in that case at least also the rectifier would need to be replaced with a new one with
long wire leads and it would be neccessary to make sure that while changing the rectifier, that it is
not mounted direct at the board but with some distance of about 1,5 cm´s to 2 cm´s "above the board.
( at the moment i think we should recommend this anyhow to get better protection added to the board to
prevent the "Big Caps" from loading up that high to dangerous voltages. It seems that the engineers at
Mountain computer just "forgot" to think about such a protection .....

Next thing would also anyhow be a check too the "Big CAps", if they have been damaged by that voltage. By specs they are 40 Volt typ and 60 Volt is far beyond their specs...
i fear that they have been damaged too .....
Even if they still now are in working condition, i believe the caps to be damaged that far, that they might become faulty very soon - due to the stress they had to carry by loading up to 60 Volts..... and if such guys start leaking they cause a lot of mess...

----------------------------------------------------------------------------------------

so now to the topic of replacement ....

FIRST OF ALL => THIS IS NOT CONCRET PROPOSAL FOR SPECIFIC CHIP !!!!!
It´s far more e result from search in my archives of datasheets searching for an IC that can handle
voltages above 40 Volt ! AND YES - I KNOW THE LT1083 and the LT1085 to be NOT PIN COMPATIBLE AND
THAT THE NORMAL CHIPS ARE ADJUSTABLE REGULATORS !!!! BUT if the entire stuff is not read - THE VALUABLE
INFORMATION IS NOT RECOGNIZED => So first read the whole stuff before jumping to the reply button !
Otherwise you will not be able to answer to the really given questions - but instead just fill the thread with useless comments !

So please take a look to the first page and then READ THE COMMENT BELOW THE PICTURE :

IMPORTANT INFO´S ARE MARKED VIOLETT!
Please before yelling - read the Features list and put your eyemark to 8th point of the list:

"FIXED VERSION AVAILIABLE"

FIRST QUESTION: IS THERE A VERSION for 24 Volt positive availiable ?
SECOND QUSTION: DOES ANYBODY HAVE DATASHEET ON THAT FIXED VERSIONS ?
THIRD QUESTION: DOES ANYBODY HAVE KNOWLEDGE OF A SOURCE where that fixed versions could be ordered ?
REMARK : PLEASE DON`T "BOMB" US WITH GENERAL SUGGESTIONS ! we know the adresses of common
suppliers... I have emailed them with bad results ! So please only answer that question if you
know the shop to have DEFINITLY THE CHIPS "on the shelf" because you have seen them there or
because the shopowner has confirmed this to you !
Then take a look to "Applications":
At this page is stated that the LT1085 version can handle 3 Ampere and the LT1083 can handle
up to 7,5 Ampere. BUT ! => In a later page it is shown in the tables that the power delivered
is dependent to the difference between Inputvoltage and Outputvoltage !
If the difference is large ( and this canhappen in this box ! ) then it turns out that only the
LT1083 Version will be useable in the box !
Then take a look to the "Typical Application"
The resistors are only needed with the ADJUSTABLE VERSION !!!
IF TERE IS A FIXED VERSION as claimed in the upper intro at features -
THEN THERE SHOULD BE NO NEED of the resistors and THE "ADJ"-pin would turn to become a "GND" pin by
replacing the ground point of the 365Ω Trimmer!

So now lets take a view to the second page:

AGAIN THE IMPORTANT PARTS ARE MARKED UP VIOLETT AND the violett text in the page!
So lets just lift the eyes to the top of the page to the "Maximum Ratings"
If we could get the "obsolete" typ within the TO3 case ....
then the maximum Inputvoltage would result from the ability of the higher difference of 35 Volt to:
35 Volt (permitted Difference ) + 24 Volt ( output voltage ) to = 59 Volt (permitted input voltage ) !!!
and besides the IC in the TO3 case permits larger range of operating temperature
AND THAT IS IMPORTANT TOO because:
within the box there is only limited space for cooling sink and that sink is not the strongest !
So permitting larger range of temperature enhances us from the need of search for "special cooling sink!

The fact that the TO3 case is marked to be "obsolete" means that they don´t make that chip anymore -
BUT THIS DOES NOT MEAN: YOU CAN´T GET THAT CHIP! It means: You have to search for such kind of chip
and if you find a shop, that has them - then you should immediatly buy 2 of them because otherwise
in later future you should have a replacement on your own shelf !

So HERE WE ARE BACK TO QUESTION 3 from above:
THIRD QUESTION: DOES ANYBODY HAVE KNOWLEDGE OF A SOURCE where that fixed versions could be ordered ?
REMARK : PLEASE DON`T "BOMB" US WITH GENERAL SUGGESTIONS ! we know the adresses of common
suppliers... I have emailed them with bad results ! So please only answer that question if you
know the shop to have DEFINITLY THE CHIPS "on the shelf" because you have seen them there or
because the shopowner has confirmed this to you !

AND JUST AGAIN THE REMARK : I KNOW THAT THIS CASE IS NOT PIN-COMPATIBLE !!!
But if this would be the only reliable solution - and only if this turns out - and only if we find a source for that chip - THEN IT MIGHT TURN OUT a problen to think about !
At the moment we don´t know : Is fixed version availiable ? Is it in TO3 case availiable ?
so i regret discussing about the color of the eggs that the bird hasn´t placed in the nest yet.....

now lets take a view of page 3:

again the important parts are marked violett:
So HERE WE ARE BACK TO QUESTION 3 from above:
THIRD QUESTION: DOES ANYBODY HAVE KNOWLEDGE OF A SOURCE where that fixed versions could be ordered ?
REMARK : PLEASE DON`T "BOMB" US WITH GENERAL SUGGESTIONS ! we know the adresses of common
suppliers... I have emailed them with bad results ! So please only answer that question if you
know the shop to have DEFINITLY THE CHIPS "on the shelf" because you have seen them there or
because the shopowner has confirmed this to you !

and now lets take a look to page 4:

again the important parts are marked violett and violett text added:
this page contains the definitions that cause my explenations to the earlier pages:
TO3 case is better than Plasic case -
and
- because of the decreasing power in Amper dependent to the difference of inputvoltage and outputvoltage
it is upmost important to search for the fixed version for 24 Volts of the typ LT1083
because the LT1085 only might deliver 0,5 Ampere at "high voltage" input due to internal restrictions...
while the LT1083 will deliver at the same difference at least 1 Ampere...

so lets take finally a short view to page 5:

again the parts marked up violett are important to us:
This is the page with the bad news....
it displays the "obsolete" "M"-typ in the TO3 package only at the LT1085 - which is the "weak" 3 Ampere version !
At the stronger LT1083 that handles up to 7,5 Ampere that case is NOT LISTED.....

well just as private remark: i dont understand the policy of the company that equips the "weaker" IC with
the stronger Case with better power consumption..... BY LOGIC IT SHOULD BE THE OTHER WAY ROUND.....
Under normal conditions the stronger chip will have the higher powerconsumption and therfor demand the stonger case.....

so anyhow - and this might turn out - if the researches don´t turn back with concrete results for souce availiablity and availiability for fixed typ of 24 Volt the whole thing might turn out to be just academic discussion......

BUT to turn back to non academic discussion:

There might turn out by measurements that still there is demand for a fixed Voltage regulator
that permits "high Voltage" input above the 40 Volt limit !

It would be at least a recommendation to take such a replacement
- after the results seen in the measurements above !

The problem is: i know they existed years ago - because i have realized several applications
for laboratory measurement systems with such chips....
but now searching for the data after several years it turns out that such chips are not common anymore....
i did not find any of the "ancient" datasheets ....
but i do remember that it was also in those days extreme difficult to find such chips and
it turned out that they had been only used in military applications in normel condition....
The chips had the specs:
24 Volt fixed voltage 3 Ampere or 5 Ampere in TO3 case, max. inputvoltage 60 Volts !
12 Volt fixed voltage 3 Ampere or 5 Ampere in TO3 case, max. inputvoltage 60 Volts !
and as far as i remember they had special marking - something like 78GKC24 or 78HV24.....

if anybody has such a device on his shelf ... please be so kind and leave here feedback:
which labeling and which company ?

sincerely speedyG

My current line of thinking on the source of the 60V at the caps:

1- It's possible the 60V is coming from the other half of the power supply (downstream of
the regulator). A partly damaged regulator could be allowing the other parts of the supply
to pump up the voltage.

2- Another possibility is the Bridge Rectifier is bad. Easy 10 minute testing.

3- AC transformer brick isn't fully isolated. Let's test that by seeing if there is
isolation between the input and output. Easy 2 minute testing.

4- A ground point is open somewhere.

5- Something is wrong in the oscillator output stage (the 2 fets that swing the
transformer).

Further testing and measurements will narrow this down and eliminate suspects.

It would be nice to be able to gradually power up the circuit with a variable supply instead
of blasting it with the full output of the powerbrick. This way we could watch for any parts
getting stressed early on and kill the power immediately before the magic smoke escapes.

I'm thinking about some more passive measurements with the regulator out of circuit. How
much resistance does the regulator have to push into (the rest of the supply) in this case?

Are there any shorts in the supply that could have overloaded the frontmost parts like the
regulator and the bridge?

It is vital to ask and determine why a part was damaged before replacing.

Hello Keatah,
i agree...
when the current measurements have been completed by Steven, this will become next task, if that 60 Volts remain.
Otherwise i believe we can consider the damaged IC to be the "pump".

What really worries me is - that even if the IC is replaced by a working regulator - there is no real protection against this kind of loading up.... therefor i propose it to be a good idea, to add protection by a fixed resistor at the rectifier, that protects the board against such kind of accidents.

Such a protection would also be good to protect the IC and ensure that the voltage won´t exceed the limits of the inputvoltage anymore... and that would make search for replacement of the voltage regulator more easy.

sincerely speedyG

@Speedy and all:
Yes. If the 60 volts remains, then Bridge Rectifier is blown up. Or on its way to complete
fail.

A failed regulator may (or may not) be the actual pump itself, but it could let the swinging
FETs' action back into the capacitors. We don't know yet. More measurements are needed.

If we add a resistor to limit voltage build-up, (which I don't want to do) it is important
to pick the right size so that can limit voltage build-up while at the same time not loading
down the power brick and bridge. I feel a Zener and a multi-watt power resistor from the +
of the big caps to ground would work better. I even see this as a safety feature. And
wouldn't mind adding it to my junk chassis board.

I would rather design a circuit so that once it detects a voltage level, say 39 in this
situation, it causes a current surge big enough to blow the fuse and disconnect the power.
Or just use a self-healing fuse. Or better up-front current limiting. There are many
possibilities. Transorbs. Zeners. TVS. All kinds of things! Take your pick!

On the other hand, I'm not too worried about fixing this high voltage situation & loading
up. It isn't an every day operational condition. We're working with failed board and thus by
default have unusual conditions present.

Once we get into redesigning things where do we stop? I could certainly improve upon the
original designer's intentions and results. I could build in all sorts of indicators and
controls and what not. Feedback loops and limiters 'et al. And may do just that. It is my
intent to determine the cause of failure here and repair it back to functional status.
Engineering improvements can then be thought out in a relaxing and enjoyable manner at
leisure.

In fact I wouldn't mind donating my junk chassis board just for that purpose. Maybe put some
bus indicators on it, a beefed-up supply, classic red-led 7-segment voltage displays. Some
toggle switches. A fan. Maybe change the slot switching logic. There's lots of room inside
these. And it is possible to cut out the entire power section with a Dremel tool to make
more additional space! That sort of thing.

It is in my personal experience to not over-rate or beef-up a component too much. I'd rather
have the regulator act as a secondary fuse and sacrifice itself rather than toughing-it-out
and staying operational and blowing up 10 parts downstream. A graceful failure.

Right now the NTE1924 is a $10.00 part and a drop-in replacement/equivalent. And I recommend
we use that. It's available, it's cheap, and since we can't "ramp-up" the supply gently, I'd
rather blow this $10.00 part again as opposed to a $50.00 part. Or worse, maybe the $50 part
would hold up and blow other things between it and the true (yet to discovered) failed part.

The caps?? Are they bad?? Who's to say without dielectric breakdown and discharge curve
testing and all that. My experience has shown them to be tougher than one would think. They
haven't exploded yet, so thats a good sign. But as always your mileage may vary.

And speaking of modifications, I was wondering about getting these heatsinks for the fets --
http://www.aavid.com/products/standard/pf527g -- These are simply bolt-on and wouldn't
affect the originality or "authenticity" of the chassis. A 100% reversible modification. Put
some bling lights in it. Make it look ominous and cool.. yeh ok..

@Steven and all:
The next steps should be making a few more tests & measurements as described above. Meantime
we can order the NTE1924 and have it ready to go.

Hello AFmembers,
hello Steven and Keatah,
in wide ranges i agree to the above said...

My only proposal would be to wait with ordering the NTE1924 one or two days
to first get the measurements on the big caps after the SG340-24 has been removed.
If the measured voltage then remains below 40 Volts, it gets time to order.
the reason: If the 60 Volt don´t drop it might become useless to replace the Q8 with a
IC that is limited to 40 Volt inputvoltage and could blow up immediatly again after assembling it.

The adding of a coolingtopsink as proposed would be anyhow a good idea. The heatsink on the board
is rather weak and the adding of such additional coolingmethod is good investment in lifeprotection of the Q8.

sincerely speedyG