My CTC-16XL Project.

[Yamamaya42]

I will start hitting that area this weekend, meanwhile, I took the removed caps to work and checked them with a calibrated capacitor tester , it does not test leakage , but this is the result.

c98 .01 = .005
c99 .01 = .013
c100 .001 = .001
c101 330pf = 298pf
c102 330pf = 342pf
c103 330pf = 351 pf
c104 330pf = 339 pf
I prob did not have to replace some of them, :/ but ya can't test them in circuit, and I can't sneak this tester home! :p

[Yamamaya42]

Originally Posted by old_tv_nut

Rolling up or down shows that you are changing the frequency, but if you can't get it to slow down to zero beat with point J grounded, there is something wrong in the oscillator (or maybe control) circuit. The caps you replaced would not be critical unless completely dead or very wrong values. Look for trouble with the cap values in the oscillator and control circuits.

C112, C118, C114, C111, C108. Due to C108, there should be no 3.58 MHz energy whatsoever at point J, only the low frequency beat note.

I can see what you mean with C108, high freq filter, not sure what C109 does with it, orange drop through r167, as I mentioned, I intend to go over them this weekend.
As you can see above, c98 was WAY out of tolerance (low) for 20% C99 just barely inside (high)
C100 OK , C101-104 2 within , 2 out of tol for %5...
I just hope that this won't be the same for all caps in the set

[old_tv_nut]

C108, R167, and C109 in combination with the output impedance of the phase detector (two 1 Megohm resistors in parallel) form a classic phase lock loop filter. At frequencies below 48 Hz (the combination of R167 and C109), the frequency response is rolling off from 1.0 at DC to 1/15 above 48 Hz (33k/500k). The response is flat up to about 480 Hz (the combination of R167 and C108) and rolls off again for frequencies above 480 Hz.

The bandwidth and transient response of the closed loop (and the amount of phase error due to frequency drift with temperature) is also very much dependent on the loop gain, which involves three factors: the gain of the phase detector (volts per radian); the gain of the control tube (picofarads per volt); and the gain of the oscillator response (radians per second per picofarad). [radians per second = 2 times pi times Hz]

To design this circuit, you would measure/design the gains for each of these stages individually, and calculate the RC filter values to give the desired frequency and transient response (damping factor).

[Yamamaya42]

just an " out there " question, the Sams says r164/r165 as matched pair, same with 149/150, right now, the person who worked on it before me has 5% carbon film ones in there, i only checked by looking that they were correct value, how critical is the " matched pair" ? should I bother replacing them with 1% metal film ones I have ( checking that they are as close as they can be to each other via a calibrated VOM ) or just leave it as it is now with the 5% ones in? i have no idea if this person checked for them to be a " match " or not, or does it not really matter with 5%?

[old_tv_nut]

Looking at the photo of the color board, it appears the original resistors were carbon composition. I think these could change value by a couple of percent due to heat of soldering, subsequent moisture absorption, etc. even if initially matched. However, it's hard to guess what they had in mind since they don't tell you.

If you just throw in 5% tolerance resistors, the center voltage could vary by as much as +/- 1 volt (because there is 40 volts across them), so that must be too much, since unmatched resistors are not allowed here. 1% resistors would cut this down to +/- 0.2 volts, which certainly must be close enough, since this is about the closest tolerance you could achieve with an analog ohmmeter anyway. Since carbon comp resistors probably would not hold such a tight tolerance, my guess is that somewhere in between, like a 2% match, is OK.

If this were a professional monitor circuit, it likely would have a pot between the two fixed resistors, with a procedure for zeroing the output voltage under certain test conditions.

[old_tv_nut]

By the way, it lools to me like DC is blocked everywhere in the phase detector circuit, so you could measure these two resistors without unsoldering them.

No such luck with the color killer detector, because of the DC path through R147.

[Yamamaya42]

Originally Posted by old_tv_nut

Looking at the photo of the color board, it appears the original resistors were carbon composition. I think these could change value by a couple of percent due to heat of soldering, subsequent moisture absorption, etc. even if initially matched. However, it's hard to guess what they had in mind since they don't tell you.

If you just throw in 5% tolerance resistors, the center voltage could vary by as much as +/- 1 volt (because there is 40 volts across them), so that must be too much, since unmatched resistors are not allowed here. 1% resistors would cut this down to +/- 0.2 volts, which certainly must be close enough, since this is about the closest tolerance you could achieve with an analog ohmmeter anyway. Since carbon comp resistors probably would not hold such a tight tolerance, my guess is that somewhere in between, like a 2% match, is OK.

If this were a professional monitor circuit, it likely would have a pot between the two fixed resistors, with a procedure for zeroing the output voltage under certain test conditions.

I will most likely just put the 1% ones in that I have, to be on the safe side. You can get .5%, but that's prob overkill for this situation.
mouser # 71-CMF551M0000DHR6
And even though they are 4 hours from me, It cost more to ship it than the parts themselves, not worth it. :p

[Yamamaya42]

Set color killer control fully CCW, set Tint control to center of it's range.
-notes killer in full off setting to be set later, Tint mid range to be calibrated later.

Connect color bar generator to ant inputs, adjust for normal color reception. Short pin 1 of Burst Amp, V22 to gnd.

notes, suggestion of color bar generator was for non-changing stable input, and any paused video from DVD or the like can be used, or else levels will change as video moves, changes, Shorting of burst amp pin 1 is to be sure no killer input is going to it & it is in fully on state. Easy access to pin 1 via killer pot R10.

Connect DC probe of volt meter to pin 1 of phase detector through a 470k resistor. (V23)
adjust A15 for max deflection ( most voltage ? ) if no reading, osc is not running, adjust A16 to start, & retry. Remove short from pin 1 of burst amp, adjust A17 for max deflection on meter, make sure osc is running and locked in.

notes , Phase detector looks to be isolated , via the 330 pf caps and fed pulses from L30, so I assume the 470k resistor is used to not load it down when doing this setup, I did see the voltage peak when adjusting A15 & A17 , but no kind of “lock” per-sey, just different degrees of barber poling when the short was removed from the burst amp, but it DID have some visual effect when removed.

Short point J to GND, remove meter from prev step, adjust A16 until color bars stand still or drift slowly. Remove short from point J , check to see that color bars sync with low input signal. Retouch A16 for best hold as needed.

notes, I tried this step, even though I was never able to get any sort of “lock/hold” as mentioned in the step before, and no adjustment of A16 was able to get a hold, just as mentioned above, varying degrees of barber poling.
I have not gone on to the last steps, obviously.

I did see healthy pulses into the phase det, ( shown in prev entry ) , and at point J, a sign-wave, at 1 ms per division on the scope, but could also see noise at higher scan rates, ( not sure if that's normal or not, prob is. )

I am still not done going over the osc / osc-cont area yet cap wise, not really found much yet, but SOMETHING, is not right, cause with the fed in video, going through that AFC set up, I should have been able to get a lock of some sort by now.

And as I mentioned, I know for sure that L24 & L25 were replaced before I got this thing, and never really adjusted as far as I can tell, but as mentioned a while back, this is not related to my current headache, and just hope it does not come back to bite later...

[old_tv_nut]

At this point it would be helpful if you had an accurate frequency counter and could measure the range of adjustment with L31 (A16) to verify that you are always on the low side or high side and cannot get nominal frequency. It seems that there must be something wrong with the oscillator circuit (wrong capacitor value or such).
C112, C118, or C114 would be prime suspects, or even one of the .01 bypasses (C110, C113), since you say the board was mangled by whoever had it last.

[Yamamaya42]

Originally Posted by old_tv_nut

At this point it would be helpful if you had an accurate frequency counter and could measure the range of adjustment with L31 (A16) to verify that you are always on the low side or high side and cannot get nominal frequency. It seems that there must be something wrong with the oscillator circuit (wrong capacitor value or such).
C112, C118, or C114 would be prime suspects, or even one of the .01 bypasses (C110, C113), since you say the board was mangled by whoever had it last.

all I have is the freq counter in the meter I bought, and its only 4 digits, I can try to get one with a few more. but they may be costly. by this weekend i will have gone over the whole circuit.
That guy turns out, did NOT have a new xtal, and ended refunding me so if on the odd chance it is bad, I'm not sure what to replace it with.
The bottom of the pcb has white flux residue all over it, not sure WTF type of solder / flux was used, and it is a bitch to get off, but i got a lot off around the osc area, i have no idea if it was causing problems, bit it is ugly..

[old_tv_nut]

One of these should work:
https://www.ebay.com/itm/10-x-3-5795...fe97e#shpCntId

What have you tried to remove the flux?

[Yamamaya42]

isopropyl alcohol 99, Chemtronics ES835B , and a remover for water base flux,

what ever was used, it is very hard to get off.
I can tell, that the untouched parts of the set have OLD rma flux on them still.

this i almost never used any more in production, but was used a lot back then.

we use water soluble and no-clean,( lead free) I had to get special order leaded rma solder to work on this.

when the resistors were replaced, some unknown type of solder was used, and the flux mix caused a white scale that has to be almost scraped off, not alcohol or Chemtronics would really soften it much, I gave it a good try and got some off but was afraid of damaging things.

17 pF on the crystal? that narrows things down a lot ( so may options looking them up )
i assume Mouser #: 520-HCA357-17X is the same?

[old_tv_nut]

Actually, I think Zenith used 17 pf, and RCA used something larger, like 22 pf. So search on ebay to see if you can find something like that. But if not, I think the ones cut for 17 will work with a slightly different tuning of A16. The important thing is that the crystal sees the equivalent of the capacitance it is cut for as the result of everything that is attached to it. Crystals tuned for one or the other will change the gain of the control stage, which will affect the pull-in range and transient response to some extent.

[Yamamaya42]

Mouser has crystals @ both 20pf and 30pf.
I don't know for sure that the original is bad, but I don't know that it's good either.
I do know it's oscillating, but something seems to be off with the circuit.
If I understand things correctly, with point J grounded, I should be able to hook a frequency counter to L33 ( general area ) and be able to adjust L31 (A16) to get 3.579545 mhz, and as well as a respectable range above and below.

[Electronic M]

I wonder if the door knob who worked on it before you used something like acid core plumbing solder?... You're never supposed to use that on electronics but some people did, infact if you go back to the 20s and 30s some electronics makers used it.