|
Vintage Test Gear and Workshop Equipment For discussions about vintage test gear and workshop equipment such as coil winders. |
|
Thread Tools |
22nd Jan 2020, 2:43 pm | #21 | ||
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
She's now chooching at full capacity, thanks everyone for all the help
Quote:
Quote:
|
||
22nd Jan 2020, 4:23 pm | #22 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Yeah, channel 2 is definitely a bit dodgy, on certain ranges it also shows what looks like compensation issues that it doesn't on others/on channel 1, I assume that the screw/variable capacitor corresponding to that range needs to be adjusted as well so you can compensate evenly through the ranges? no idea what screw corresponds to which range though :/
|
22nd Jan 2020, 5:08 pm | #23 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,874
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
That's correct – however, to set the attenuator compensation correctly you need a square–wave/rise time at least no worse than that of the Oscilloscope you're adjusting – A Tek.106 or the later PG506 in the appropriate main–frame is needed, set to give a 1MHz output equivalent to about 5 divisions for each attenuator. With this connected to the 'scope via a matched 50 ohm termination, the shunt compensation trimmers are adjusted for the best corners of the display, without any overshoot or ringing, while the series ones are adjusted at a lower frequency, usually! If you can't get hold of a fast–rise Oscilloscope Generator, a really good function generator will do at a pinch but the compensation wouldn't be as accurate! (For basic scopes, up to 10MHz, you could probably get away with a F.G!) In the absence of the full written 3122 Manual, look for another similar British Oscilloscope of the same bandwidth as the 3100/3122 with the same or similar attenuator arrangements, that you can get a manual for, and use the other 'scope's bookwork as a guide! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |
22nd Jan 2020, 5:18 pm | #24 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
i mean the compensation seems Waaaay off, not the kind of thing you'd need a super high precision reference to notice- seems to be overshooting a significant fraction of the size of the square wave - it's visible even with the scopes internal compensation output (other channel doesn't see it,gives a perfectly flat square wave so i doubt it's a problem with the calibrator as such).
wondering if that channel might have developed a fault tbh :/, gain is also quite noticeable- on the order of 20% or so, with channel 1 seeming to correspond far better to the numbers listed on the calibrator at least |
22nd Jan 2020, 5:40 pm | #25 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
ach, no compensation problem- the selector switch for channel 2 is just a bit dodgy( i can make it show the same by applying a little pressure to the rotary selector/wiggling it about till it stabilizes, and it feels a bit "squishier" than channel 1s.
gain problem still there, and actually more like 25% 4.5vs 6 divisions, not a probes issue, swapping them causes the same readings |
22nd Jan 2020, 7:04 pm | #26 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,874
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
The way to trace this type of fault is to feed an identical signal into both channels by means of a BNC "T–Piece", set the gain–switch to the same factor in both channels (I recommend 1V/div), adjust your generator for 4 div. above and below the centre, then use a second oscilloscope to see where it's getting lost, but if you've not got one, you can use a DVM with a 1uF capacitor (NOT electrolytic – some of the test points have negative voltages on them!) in series with it's red (+) lead, then feed in enough sine–wave input at 400Hz or 1kHz to give 8V pk–pk on the screen if the good side as before! You can then use the DVM to trace the signal thro' the amplifier stages, loss of gain, if not due to yakky switch contacts in the attenuator switch circuits, may indicate damaged input FETs or transistors in the early stages! I'm afraid you'll need the 3102 circuit diagrams from Bill to solve this, unless you can prove the contacts themselves are the culprits! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |
22nd Jan 2020, 7:28 pm | #27 |
Octode
Join Date: Oct 2007
Location: Exeter, Devon, UK.
Posts: 1,554
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Cossor 3100 and plug-ins 3102 YAmp and 3122 Timebase.
I am glad you have solved the Time base problem. Now you will want the Y Amplifier details. I attach the circuit diagrams for the 3102 dual trace unit, and the pages from the manual relating to setting up the attenuators. Don't be put off too much by this talk of an expensive Tek106 or similar. The built in calibrator will give a pretty effective square wave for setting the attenuators, though the faster the rise-time, the better the result. Chris 55000 suggestions seems very sensible. It is surprisingly difficult to fault find without a second oscilloscope. wme_bill |
23rd Jan 2020, 11:43 am | #28 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
o lordy, what on earth is transistor 1/10? it's some kind of to78 dual jfet package.
(7 leads with one trimmed) The circuit diagram marks it as spec. 128(?) and its metal case is marked wd111 7235? none of which give very convincing hits on google |
23rd Jan 2020, 1:38 pm | #29 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
hmm, actually, i wonder if the main contributing factor to the bandwidth, is actually just those tantalum capacitors? specifically c4/c18
if you look at the "dc balance" circuit, you'll see that twiddling the position knob on the front of the unit produces an at most ~0.2v in voltage from one extreme to another to the base of TR1B, and that's enough to more than take the trace off the edge of the entire screen, both ways, so maybe 0.01v to create 2 divisions?. it's a really big capacitor too, implying that line needs serious decoupling, as the frequency increases, isn't the ability of that cap to decouple going to decrease a lot due to ESL, if that line needs to be stable to within <0.01v ripple under load to stay within 2 divisions,that tantalum might not be capable of that at higher frequencies? wouldn't take a huge parasitic impedance to cause that, something on the order of 10ohms? dunno if that's how it actually works just spitballing , and i don't actually have equipment good enough to test the frequency response of the scope, so just theoretical for me. |
23rd Jan 2020, 2:44 pm | #30 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
edit, math wrong, probably an ESR change < 5ohms would do it, maybe as low as 2,
that seems reasonable thing to hit at about 35mhz? idea being - jfet turns on amount due to the voltage difference between GS, the jfet turning on turns other transistors on, which pulls the voltage of the Jfet source down, which acts to lessen Vgs, which turns the jfet less, which turns on the following transistor less, and acts to oppose the amplification. Yeah, i think that power line needs to be as stable as possible... |
23rd Jan 2020, 3:28 pm | #31 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
I mean impedence not esr, sorry.
to get a 1ohm impedance at 35mhz requires ~5nH, not taking into account trace impedance and the such. with all that, more than 10nH doesn't seem implausible, causing several ohms worth of impedance? Forgot to take into account the DC balance, but it still applies - The capacitor is connected to both, so the ripple would effect both roughly equally, and there's a differential amplifier, so the difference would change by a similar factor? someone who's better at analog circuitry than me help me out here XD Last edited by golexica; 23rd Jan 2020 at 3:54 pm. |
23rd Jan 2020, 5:33 pm | #32 | |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Quote:
I'm a little confused what the capacitor is for here? wouldn't 1uf be almost transparent at 1khz?, reactance of 160ohms, isn't that too low?, at least compared to the input of the multimeter, which should be 1Mohm or so on voltage? Last edited by golexica; 23rd Jan 2020 at 5:46 pm. |
|
23rd Jan 2020, 7:16 pm | #33 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,874
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
H!
Many of the inexpensive DVMs will give nonsense results or go haywire on d.c. with an a.c. component superimposed on it, so the purpose of the 1uF capacitor I suggested is to block the d.c. potentials w.r.t chassis that will be present in the amplifier circuit! Compared with the impedance of the meter input, the reactance of 1uF at 400Hz or 1kHz will be negligible, so the meter (switched to a.c.) should show a reasonable representation of the signal voltage! Chris Williams PS! Many of the much more expensive Flukes, HP/Agilents, Keithleys, Siglents, etc., (mostly bench instruments) do have an (AC + DC) function that allows you to select each component and display them separately, but I write my fault–finding notes for use by members who may only have basic instruments!
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! Last edited by Chris55000; 23rd Jan 2020 at 7:27 pm. |
23rd Jan 2020, 7:59 pm | #34 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Mystery solved , probing across 7 and 8 ( set gain for ch1) gives 162 ohms (250 pot), probing across the equivalent for ch2 (24/25) gives...294, on a 250 pot.
how did it get so far out of wack? looking at the adjustment pot, it seems to be attached to a long shaft...that goes through the middle of the shaft for the range selector,with a tiny screw in the middle of the rotary selector, that you use to turn that shaft, The ickyness on the rotary selector, must have caused a snag on the inner shaft, messing with the range calibration. That seems like an unnecessarily complicated design, how often do you really need to adjust the calibration pots? well at least i know where all the position calibrators are now |
23rd Jan 2020, 11:24 pm | #35 |
Nonode
Join Date: Dec 2012
Location: Kirk Michael, Isle of Man
Posts: 2,350
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
WD111, 7325. This sounds like a dual FET, type WD111. The 7325 is week 35 of 1973, prodn. date od component. Many of these FETs and Dual Fets had in house numberring. many had UC, meaning manufactured by Union Carbide. I don't have any data right now, but the WD111 does seem to ring a bell.
Les. |
28th Jan 2020, 4:17 pm | #36 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
urgh, a doozy, mean y plate voltage was too low, and the voltage drop over an input resistor to the y amplifier (R35) was
too damn high ~13.5 half volts on a 1w 150 ohm, and it would have been derated because of proximity to several really hot components, pcb around it was brown, and it was starting to drip... problem seems to have been the darlington tr11- was showing a diode between emitter and collector, which is not supposed to be there? and replacing it seems to have gotten the y plate potential up to 65, and the voltage over the input resistor down to a more reasonble 11 ish. what confuses me a little is c13 ( a trimmer cap ), it looks unhappy/darkened compared to the other, which is weird, because it has 200 ohms of parallel resistance, so a short between the plates seems unlikely, it does have a "shadow" so maybe just radiant heat? shadow doesn't seem directed at anything particularly close by, and nothing that seems to show signs of having gotten very hot? maybe resistor 38, but it's a stretch. also, there aren't any actual ratings for these trimmer caps in the parts list, so i've not idea what to replace it with? More to the point, i don't even know what it's for?, or what it's supposed to be adjusted too? maybe the HF compensation mentioned previously in the thread? I guess I'll leave it in place for now, seems to be working as is. |
28th Jan 2020, 5:04 pm | #37 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,874
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
It's a common type of PCB mount plastic trimmer you can buy, still made, 2–20p should be suitable! The reason for its brown, discoloured appearance is simply heat rising from the transistors & other components over a long period of time! One of these'll be fine to replace it! https://rover.ebay.com/rover/0/0/0?m...2F123069500502 If you can't get hold of the full instructions, feed in a square wave of about 1 MHz from the best generator you have (or the calibrator o/p of the 'scope) and adjust C13 for the best quality corner display you can get with the minimum of overshoot, etc., and it won't be too far out! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! Last edited by Chris55000; 28th Jan 2020 at 5:18 pm. |
28th Jan 2020, 5:08 pm | #38 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
i'd seen those, I was just worried about voltage/ temperature rating, i don't imagine the voltage getting high, but the temperature is evidentially scorching
|
31st Jan 2020, 3:44 pm | #39 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
I think i'm approaching the point where it's "good enough", and besides, i'd probably need a second oscilloscope to bugfix parts of it: blanking voltage at 19 on pcb j seems to not want to rise above 75v, and manual says 90v, and i can't find any component that's obviously broken, and i can't use the oscilloscope to see what's up...because the oscilloscope is blanked during blanking .
one thing that i'm unsure about is the "chop" mode- if i turn the brightness way up so i can see the transition between the two traces. Is it usually that noisy/overshooting? it would normally be blanked out and invisible at normal intensity levels so it's not necessarily that big of a deal, it just seems excessive? the chop circuit seems to be based on two nand gates with their outputs wired to the other's input (a flip flop?, but one that's astable?) There's also a phase offset between x and y channels?(square wave test output fed to x and y produces a hysteresis style loop) there must be a way to adjust that? Last edited by golexica; 31st Jan 2020 at 4:08 pm. |
31st Jan 2020, 4:26 pm | #40 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
|