Hi chaps,
Back again, spent some time at the workbench... so go grab a cup of tea and some biscuits to go with it... as you have figured by now, I always struggle to keep things short, no matter how hard I try.
Point noted for the bridge rectifier, rectifiers at large... and bridges in old Hamegs too, because I happen to have one...which got to get some use today BTW, to assist in the repair of the Tek.
In short : as planned, I bypassed th regulator and powered the inverter from my bench power supply. The scope is back to life, tested it without the digital board and as a 100% analog scope, it works fine... or almost, see further down.
But before I order the parts to fix the pre-regulator and reassemble the scope, there are still 3 points that concern me, I could definitely do with your input on the subject.
For clarity, let's unfold things in the order :
First, before I took the risk of powering the inverter and the entire scope with it, I :
- made some space to get access to the shunt resistor and SCR : removed the tall heat sink as well as the cooling fan mounting piece.
- cleaned the mess/carbon deposits from the blast, to better/more accurately assess the actual extent of the damage. Some IPA with a ESD safe brush, helped. As I thought, the little T092 tranny, and resistor that looked carbonized, next to the TL594 chip that blew, were fine : they were simply covered with smoke because they so happened to live there, but cleaned up just fine and look sparkling new again. Obviously, I will check those two components electrically anyway, just in case...
- removed the plastic plate on the bottom side of the PCB, to get access to the solder side. Found some obvious signs of rework in there, old flux that did not get cleaned. But I guess that's to be expected in any old gear...
- removed the shunt resistor that blew, so I can check check for damage on the PCB underneath the resistor, because it looked like it took quite a hit !
Not sure it shows well in the pictures, but the PCB survived. The fat track of the 42V rail probably got scared to death, but it is intact, and the PCB got burnt enough for the fiberglass weaving to be exposed, but other than looking scary, there is no actual show stopper.
- Electrolytic caps : I checked the big 1000uF one that filters the 42V rail, because obviously if a cap had a good excuse to blow, this was the one. No visual sign of damage, so just measured it : no short that's a good start, and it's value is still well within spec, no worries there. So at the least I could assume there was a good chance that it did survive the blast, ie the crow bar did its job, along with the shunt resistor acting as a fuse and protecting it too.
- SCR/Thyristor: removed it from the board so I can check it properly/out of the circuit, and also so I could manage to read its part number (impossible other wise, as it's stuffed right into a corner of the chassis, facing towards the metal work.... ) because the parts list says it's an SCR2117 or something, but Google can't find any data sheet for this. Actual markings on the beast are " MTR 72 ". At least this one is known, but obsolete "obviously".. so I will have to figure out a replacement for that as well. Shouldn't be too difficult though, I guess...
Anyway, the forced and unfortunate absence of the SCR on board, would not keep me from powering up the scope, since it is powered by the bench power supply therefore safe/current and voltage limited.
- mains bridge rectifier : checked all of its internal diodes, all good, and a consistent 0.45V or so forward drop which is fine by me.
Second, I set the bench power supply exactly to 42.8V as indicated in the annotated schematics, and set the current limit to 2A as you said.
Then I crossed fingers that nothing would blow again.... victory !
The scope powered up, it was not shorting the power supply. current draw was 750mA then after warming up settled at 800mA. Scope was powered like this for many hours while I was testing it, and no incident to report.
So, followed you advice : tested the thing to make sure all was well, before rebuilding the pre-regulator. This is were I found 3 problems which could do with your input... Other the these, the scope, in pur analog form for now, works just fine.
1) Odd : the probe compensation signal goes NEGATIVE ?! Frequency, amplitude and signat shape are spot on, but instead of going from 0 to +0.5V, it goes from 0 to MINUS 0.5V ... go figure.
2) The time base is off, and by a long shot ! Error is measured at about 4.8 to 1 . At first I though well that's all right, it's just the big know that got loose again and missed a couple steps .. but not, not that simple : knob is still firmly attached/screwed, turned it all the way from lock to lock, the dial goes where it's supposed to go... no mismatch. Plus, had the knob been out of adjustment, it would imply an error fitting the usual 1-2-5 sequence.. but here I witness 4.8, too far off "5" anyway.
2 bis) The little "CAL" knob of the time base, woks only partly : The "s10" magnification works. however if I turn the knob CCW to go out of cal on purpose, it has absolutely zero effect on the trace on the screen. Yes, yes... when I say the time base off, it's obviously when the little CAL know is set fully CW, in the calibrated position...
The CAL feature has a range of 10 : 1 isn't it. So it not working, and the fact that the error I get fall inside that range... it's reasonable to assume that the time base subsystem is fine and that only the var cal feature is misbehaving, it is somehow stuck solid on 4.8 : 1 ...
Had a quick look at it. That knob has its own dedicated switch assembly, which is also mounted on a small dedicated PCB. So that's a starting point... There are only 2 cables going to this little board : one which goes to the digital board, which obviously is disconnected at the moment, and another smaller, 3 way connector which goes to the main/analog board, but this is one firmly secured in place. So I fear it's not as simple as a disconnected connector... and probing in that area doesn't look like fun, access looks shitty at best...
3) Power rails : they all look fine except for the 5V rails which has me very concerned ! Exhaustive list of my measurements follows. Measurements taken with a fixed x10 probe sorry (not ideal to measure tiny ripple...), as my slower x1/x10 probe is out of order... need to order new ones... So, x10, and measured on the Hameg scope whose lowest setting is 5mV/DIV not 2mV like the Tek. So that gives us 50mV per division. Still good enough to get a ballpark figure though....
+ 30V (CRT) : 30.3 V. Ripple : 40mVpp, 50 tops. No random garbage, but rather a perfectly clean wave form fro the inverter. The trace as jumping around a little bit but I the period was about 45us, ie 22.2kHz... and the waveforms the service manual provides for the inverter, indicate around 21.8kHz, so not bad.
+100V (CRT) : 102.8 V. Ripple : same as above (clean, inverter residue, same amplitude, and frequency as well obviously), except that somehow the waveform/pattern is different. See below for pics of each of these two waveforms.
+8.6V : +8.75V. Ripple: even at 5mV/DIV, it's as flat as it gets, no discernible ripple whatsoever.
- 8.6V : -8.7V. Ripple : none visible.
-5V : -5.8V Ripple: none visible
+5V: +5.46V Ripple: none visible.
Not 100% sure why the 30V and 100V are the only ones to display some ripple, whereas the other rails are really really flat. Probably because the 30/100 rails have fairly small filter caps, only 33uF and 270uF, whereas the other rails have much large 1000uF caps, as well as LC filtering.
But obviously my big problem is with this +5V rail ! +5.46V, what the ?!
At first I though nothing of it, thinking well who cares, there will be regulators down stream... then realized that 0.46V isn't much headroom for a regulator, even an LDO type which probably wasn't really widespread back then anyway. So, I started to get scared.... could the digital chips be getting 5.46 instead of 5.0V ?!
So, checked the "Power Distribution" schematic page, and that got me real scared indeed : I do spot a couple 78/7905 regulators, but this is only local to the "Timing" board. What the schematics shows is that indeed, the is a bunch of TTL logic chips which powered straight from that 5.46 rail ! NOOOO !
OK TTL is specified at 5.0 +/- 5%, but that only gets us to 5.25V, we are way above what is already the maximum tolerated ! :-/
I thought... OK maybe the schematics says 74 "LS", but maybe it's just a generic way of saying things... maybe the actual chips on the board, are some other family which can cope with a wider range of supply voltages. But no... checked on the board, the chips are indeed bog standard 74 "LS", so 5,25V max it is !
Then I thought... OK this makes no sense, you must chasing a red herring again.... so just to be sure, I measured the voltage directly directly on the two power pins of each and every one of these TTL chips... and sure as hell they all get 5.46V :-(
I didn't give up : could still be chasing a red herring, maybe the multimeter is somehow off ? Unlikely especially since I checked it when I bought it, and it reads fine on the other rails.... but tried a different meter anyway just in case... no joy, still 5.46V.
I don't know how the scope can still work "fine" (from a user point of view at least, bar my time base issue...), with all the TTL chips being pushed/stressed that much... but the point is, regardless of the fact that they happen to somehow function at that voltage, it can't possibly be DESIGNED to be 5.46... So now the question is... what is wrong... :-/
From the top of my head :
- partially shorted 5V secondary windings (internally/wire insulation problem) : I guess that would make for a voltage lower than expected, not higher... so unlikely
Then we could go upstream (primary winding, then maybe a problem with the inverter).. but that would affect all the rails... and they are all perfectly fine, only the 5V ones are way out.
Oh, just an idea : maybe the blast DID do some damage to some track after all, and maybe the voltage from one of the other rails, of higher voltage, is "leaking" its way to the 5V rails a bit ? No, stupid... the damage happened way upstream in the regulator, when the various rails don't even exist... so really, I am clueless !
So.... I welcome your input on the +5.46 V rail, and the preceding two problems as well !
Now picture time.
Home
Help
Search
Login
Register
