|
Vintage Test Gear and Workshop Equipment For discussions about vintage test gear and workshop equipment such as coil winders. |
|
Thread Tools |
31st Jan 2020, 7:12 pm | #41 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,853
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
The only way you'd get an identical response between X and Y amplifiers is in oscilloscopes fitted with true XY mode facilities! Earlier models like the 3100 & CDU150 had lower bandwidth X amplifier circuits compared with those for the Y channel, so the phase–response and rise–time will be different, so the answer is, no the phase difference between X & Y isn't normally adjustable! I'm not sure about the quality of the chop multivibrator drive waveform, but TTL does have a tendency to "current–spike", because two transistors in the output circuit of each gate conduct together momentarily when the o/p changes state, and this causes a transient dip on the PSU rail that (I suspect!) gets translated into the overshoot you see by the time it reaches the Y plates! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |
3rd Feb 2020, 9:27 pm | #42 | |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Quote:
I'd like to make sure the trigger and such are fully calibrated, and the sawtooth sweep seems to be "biased" slightly in one direction, which means i have to rezero the trace every time i go between Xamp and normal/delayed operation, which gets slightly tiresome to have to do constantly. This scope was in heavy need of some TLC , no idea what the original owner did to it. |
|
3rd Feb 2020, 11:44 pm | #43 | |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Quote:
|
|
4th Feb 2020, 1:00 am | #44 | |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Quote:
Interestingly, from what crude testing i could do, it seems that CH2 has a better bandwidth than 1*, implying that the bandwidth limitation is likely indeed within the dual Y amplifier * worthwhile sin/dds wave generators are expensive, but an si5351 is under a tenner, and supposedly has a rise time of something like 1ns into a 50ohm load, near the scope's bandwidth limit, most of the harmonics other than the first get attenuated heavily, and you're left with a mostly cleanish looking sin wave, and need to turn the frequency up 10mhz to get equivalent attenuation on ch2 as one- (i chose a threshold of 90% of the magnitude of the square wave, as that should convert to a 1/sqrt(2) attenuation, as in a fourier series the fundamental is 4/pi times larger than the final square wave.) obviously lot of assumptions made, but i was more interested in the relative performance of ch 1 and 2, for which it should roughly valid. |
|
5th Feb 2020, 5:39 pm | #45 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
now that my terminator arrived I did some more testing with proper termination/rise time calibration, and the difference between channel 1 and 2 is not quite as severe as before- maybe 5mhz.
something i found interesting, was that after a certain signal amplitude was reached, gain seemed to basically drop off, and the output on screen was the same size independent of the size of the input, still looked the same shape, but its amplitude didn't change. maybe implying the limitation is at the output, rather than input of a transistor? Last edited by golexica; 5th Feb 2020 at 5:48 pm. |
6th Feb 2020, 1:16 am | #46 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
not that capacitor, tried replacing with a one with lower ESR and their was absolutely no change whatsoever.
ah, well, worth a shot |
6th Feb 2020, 2:00 am | #47 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
err...is it normal to have these ferrite beads on Resistors 10 and 63 of the dual channel amplifier ?
they don't appear on the schematic, and it seems a weird place to have one- directly on the signal path, at the base of a transistor fed by only a 10ohm resistor wouldn't that by horribly attenuating? i'd love to remove it and see if it makes a difference, but am worried that it'll horribly damage the scope or something. edit: that transistor it's attached to seems very high bandwidth going off the data sheet, perhaps it's attenuating the higher frequencies there to prevent it causing issues down the line? Last edited by golexica; 6th Feb 2020 at 2:05 am. |
11th Feb 2020, 2:25 pm | #48 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,853
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
Ferrite beads, small picofarad size r.f. bypass capacitors, series or parallel damping resistors, etc., etc., are often added to series base/gate resistors or the leads of amplifying devices themselves to act as "parasitic–stoppers" in order to suppress any tendency to overshoot or ringing in the amplifier – the inference is that Cossor found the response of your particular unit to be unduly "lively" at the highest frequencies during their test and set–up procedure at their factory! Any of these found on component–leads should always be replaced where they were originally found when the unit is repaired! There might well have been a service note or bulletin in some of the manuals at one time but unfortunately such documents have a habit of disappearing without trace after the firm closes or ceases dealing with that type of T & M! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |
1st May 2020, 7:27 am | #49 |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
had time to have another look at it due to the stay at home situation, and have discovered numerous oddities.
First: the blanking amp inverter resistor R30 was getting rather hot and discolored (3122 figure 8). turns out on X amplifier mode, it's driven directly from the 10v rail for some reason. Disipating near 2w, and putting 200ma of base current through tr13 ( a tiny switching transistor that could at best pass 20ma from its collector when fully on) that's massively excessive, so I just soldered an inline resitor near wafer s5. it's super weird, they use that bare 10v line for nothing else, other than that? Second, There's a bunch of faults that do not seem like repairs, that make me wonder if this thing is a prototype or reject or something, because i can't see it passing quality control. there's wrong values of resistors in several places that were affecting the functioning of the device: two legs of a differential amplifier differing by a factor of 2, the blanking amplifier reading 65 instead of 90v because of wrong biasing resistors (failing the steps in the setup guide). |
1st May 2020, 7:55 am | #50 | |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Quote:
That seemed suspiciously low given the x amplifiers ability to produce a nice linear sweep at 0.2microseconds/division. so i looked into the front end and found the limitation. Resistor 9 on PCB B of the timebase unit, it's 1k and it's driving a coax, but that coax is short and unterminated, so it acts like a capacitor, and because of the drive strength being fairly weak, that forms a low pass filter composed of R9 and the coaxial capacitance. This is made worse by cable management instead of just running a shorter cable directly across. The proper thing would have be to amplify the signal up to a lower impedance to lessen the effect, or terminate properly and reamplify at the end (removes the capacitance, but would reduce voltage by factor ~20) that seemed complicated, so i just bodged a twisted pair directly across shortening the length and thus lessening the capacitance ( and the wire probably has a lesser capacitance anyway) It worked bandwidth is now 5.7Mhz, bandwidth's not the only important metric, but i doubt that coax was actually bringing much benefit unterminated as it was This occurs everywhere else too- drive strength to the gate out is a measly 10k, so the rise time is super slow (like 3microseconds slow) unless you terminate at 50ohms. the termination brings it to sub 10ns, but reduces the voltage signal by a factor 10k/50=200. Which is too low to be useable as a trigger unless you somehow amplify it Last edited by golexica; 1st May 2020 at 7:58 am. Reason: didn't know that was a swear word |
|
1st May 2020, 7:57 am | #51 | |
Triode
Join Date: Jan 2020
Location: London, UK.
Posts: 41
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Quote:
What level of ringing is acceptable if it get 5MHz say, or is it personal preference? |
|
2nd May 2020, 8:06 am | #52 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,853
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
Many oscilloscope service manuals do give maximum specifications for overshoot and ringing of their
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |
2nd May 2020, 8:12 am | #53 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,853
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
Many oscilloscope service manuals give specific maximum specifications for amplitude of peak–to–peak ringing or overshoot, but when you don't have the full manual or a specific amplitude is not stated, I would have thought about 0.5% of the peak–to–peak or 1/200 th of the signal voltage amplitude is a reasonable maximum! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |
19th May 2020, 8:44 am | #54 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,853
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
Hi!
I have now acquired a Cossor 3100 with 3102 Vertical and 3122 Dual T.B. plug–in units (it doesn't look like Cossor ever designed any other types), so I will be dismantling and examining it in detail with a view to repair when the time comes for it's turn – both of the plug–ins appear to have faults on them! If any Member needs any pictures or component details from my unit please PM me! I have contacted Member WME Bill about purchasing copies of the 3102/3122 books from him, and it is my intention (eventually!) to write all this info up with new diagrams and some extra information not originally provided by Cossor, such as full wiring details! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |
19th May 2020, 9:39 am | #55 |
Nonode
Join Date: Aug 2007
Location: Walsall Wood, Aldridge, Walsall, UK.
Posts: 2,853
|
Re: Cossor CRO 3100 Timebase unit repair (model #3122)
PS!
Some of the circuit data from both my uploaded 3100 main frame manual and Bill's 3102/3122 circuits is a bit "blobby" or "furry" and difficult to read in places, so me thinks new circuit & layout drawings from my sPlan software are called for, and I'll get going on these as quickly as my available leisure–time permits! Chris Williams
__________________
It's an enigma, that's what it is! This thing's not fixed because it doesn't want to be fixed! |