![]() |
Adding Composite Input (Capacitor Question)
I'm about to test inserting a composite signal into my Zenith and have a question about the capacitor orientation.
I am creating the cable to do this insertion and so far I've taken a female RCA jack and put a 75 ohm resistor across it. From videos I've seen on this, they also use a 47uf capacitor. However, the one I saw had the negative side of the coupling capacitor on the signal side. My brain tells me that since it's the signal side, that should be the positive lead of the cap. Upon reading more about coupling caps, it seems the negative side goes to the "most" negative side of the circuit. That if the signal side is lower voltage than what it's connecting to, than the negative lead is connected there. If this is true, then once I remove the IF module, I need to measure voltage on the the insertion point and if it's less than composite signal voltage, the negative goes to that side. But, if it's higher than the composite signal, the positive side goes to it. Do I have this correct? |
There is no fixed DC polarity to the composite video signal. You either determine polarity based on the injection points DC value with respect to ground. Or you use a non-polarized capacitor for the connection...Which you can make out of 2 polarized lytics of double the desired capacitance connected in anti-series...(Not parallel but rather) Series with one cap reverse polarity from how you'd connect 2 batteries in series.
|
Quote:
|
Quote:
Now my question is where are you feeding the composite signal into and is this a televison receiver or real monitor? Reason I ask is twofold. One is the possibility of a hot (AC) chassis with respect to ground and two some TV receivers don't like anything other than what comes from the video detector as I'm fighting this issue now with a Magnavox T940 |
1 Attachment(s)
I have done this but I used DC coupling into the first video amp. You're probably going to need some amplification as the set probably wants 2V P-P for full brightness. Check the schematic. I did this on a Heathkit using a MAX405 op amp and an RCA using a 2 transistor amplifier. I would go the transistor route since the MAX405 op amp is probably no longer available.
|
Yes, check for the p-p video required. In some sets it could be as much as 4 V p-p. It not only afffects the picture brightness/contrast, but also the sync circuits.
|
Quote:
This is a television, but not a hot chassis. It has a full transformer. I'm feeding it into C1 test point which is after the detector stage and where it splits off to the video amp, chroma module, horizontal module and vertical module. This test point is where you attach a scope to see the video signal from the IF. |
Quote:
|
1 Attachment(s)
Quote:
Also, tried the "new" IF module and it is either not new, or not aligned. When I installed it, the color went full and the color control wasn't functional, or it was but had no affect. |
I think I may see one issue. The output of the IF is 6.99V and the signal off that lead goes to the chroma module and that input is 6.99V. So if I remove the IF module, it drops to 4V. This where my understanding of electronics lacks. If the IF is outputting 6.99V, but the signal in the wave form is 1.4V, how is this possible? Is it possible to have the composite signal of 1.4V inside the 6.99V? If so, how does the signal not go from 1.4V to something higher?
I think it time to break out the Tek scope and look for the video signal on that test point. I tried the small digital scope I got, but it was all over the place trying to lock onto it. |
2 Attachment(s)
Okay, here is the waveform from C1 test point. I used a gated rainbow test pattern. First one is 30us and second is 20us. I used a 1x probe with 1V per division. The SM waveform says 1.4V, but this is definitely 2V p-p.
|
1) If the DC voltage is 6.99, that is the average of the video waveform, that is, the high parts of the waveform are more than 6.99 V, and the low parts (sync pulses) are less than 6.99. The difference between the max and min is 1.4.
2) The gated rainbow pattern does not go to full white level, so it's peak-to-peak range would be less than a test signal containing full whites. Does the service info say what test signal is used to get the 1.4 volts? 3) Having 2 volts p-p could be from an AGC misadjustment, but before jumping to conclusions, how does the picture look? Overly bright and unstable, or OK? |
Quote:
I believe the lack of brightness and sync was due to the IF module being removed. The chroma module input being fed from the IF module is also 6.99V and it probably couldn't compensate for the 4V with the module removed. The sync going to the video module is 6.24V and that one also didn't like the lower voltage with the IF removed. I haven't adjusted the AGC since it said to do it with a weak channel, without snow. If I do it, the signal I feed it will not be weak and the adjustment will be off, or so I believe. |
1 Attachment(s)
Partial success. I used the composite from my DVD player through a 75ohm resistor across the leads and a single 47uf cap on signal side with negative towards dvd and positive towards C1 in set. With tuner attached to IF, nothing. With tuner disconnected from IF, I got the attached. So it is now synching, but the colors are washed out and I don't know where to go from here. Am happy I got this far, though.
|
1 Attachment(s)
Very likely that first video stage is getting some bias from the IF. Not unusual.
Attached is the schematic for the video input circuit I used for an RCA CTC92. I found this circuit on the web and modified it some. The pot labeled DC sets DC level which equates to black level. I would inject the video signal after the sound trap as the sound trap rolls off everything over 4.5 MHz. Maxim used to give out free samples. But I think that's now only available to Asian EE students. Another source of video chips is Linear Technologies (now a part of Analog Devices). Not sure what their sample policy is. Something to remember, if you go the op amp route, you will most likely need a bipolar power supply. Whereas the transistor circuit requires only one. |
Image looks like my Magnavox T940 with composite input... thus the reason I'm going back to RF/IF as the Mag did not play well with the external input despite using an op-amp interface, big impedance mismatch. I used a LT-1252 op-amp in this instance.
|
I really would like to use the composite input on this set, but it'll have to wait for another day. Main reason is to see if the tuner/IF is the source of my jailbars. Pretty much check all other areas of the set. Oh well.
|
Jailbars? Are you dealing with poor quality RF modulators or front end overload as I am? My reason for the composite input on my Magnavox was due to the IF board getting damaged while I was trying to deal with some similar issues caused by the poor quality signals coming from cheap RF modulators but the composite input introduced a number of new issues related to impedance matching and the raw input levels needed to get thru the video amps, I've since gone back to RF modulation which has brought me back to square-one and I believe the solution may be the use of CATV head end RF modulators that use SAW filters. As I understand they're far more robust and have a much cleaner output signal than the cheap-o's used in home equipment, I do have experience with these as our local senior care center has gone to great lengths to keep legacy analog MATV going in the building so those with older TV's can still use them. It's a memory care type facility and many of the residents can't operate the "new fangled" flatscreens and need their old rotary tuners from long ago, just doing our best to help those in need.
As it stands right now my issues are with a limited AGC range and what appears to be some type of "saturation" that is washing out the colors on all but the strongest saturations, set chassis in question is a T940 and thanks to member Damen I was able to get another good IF board. Now it's a matter of undoing all of my composite input mods and bring the set back to original. |
Quote:
If the new IF doesn't fix the jailbars, I'm at a loss as I've replaced and tested every component that could cause the bars. |
Okay so are you saying the color & tint controls have no effect at all? I know you're working with a Zenith set but what series or model? This almost sounds like broken connections between the tint/color pots somewhere along the lines.
|
Quote:
It's a '74 Avante with the 25EC58 chassis. |
Guess we're sorta in the same boat here. I'm working with a Magnavox T940 chassis that seems to have its own ideas on how to deal with external video signals. Problem started out when I broke my IF strip board and destroyed a coil so I decided it would be easier to just inject a video signal in behind the detector diode, this introduced some impedance matching issues which were sorta solved with a simple matching network but like in your case the internals of the set don't use the "standard" video levels and the picture was weak & blotchy. Some other members pointed me to a composite input circuit used by an earlier RCA (don't recall the chassis) and I built it up on a perf board tube socket and all but it introduced a new set of challenges and was superseded by an op-amp design using an LT1252 and configured to produce up to 8 volts P-P and it worked... sorta. Other issue I was having involved sync and DC restoration which I tried to inject using a diode based circuit but the Magnavox would not play nice no matter what I did.
I bypassed the 6GH8 sync separator completely and went with an LM1881 sync sep/countdown chip from TI and configured it to feed the horizontal & vertical directly which resulted in a very stable lock no matter how much things varied but the picture quality itself was never 100%. The set has been in the family for 56 years now so I know what it's capable of... well back in the day and now no matter what I do or how I do it I can't get to that level again. Pix tube tests very good on my B&K 467 showing 1.6ma on each cathode and over the past several months I've been slowly replacing resistors and other parts that don't measure up in their tolerances so everything is what's specified on the schematic (Sams 1144 folder 1) at least academically. So recently I put out a want for a replacement IF board as I wanted to bring everything back to factory stock but with all the new parts and a fellow member pointed me in the right direction, two weeks later a carefully removed IF board and the service manual for that donor set arrived at my door. Just want to say Thanks Again to damen for helping me with this. That afternoon the IF board was carefully transplanted with the hopes everything would fall into place... then reality hit. I had to go back on all of my mods and restore the sweep board back to original then double-double check all the wiring points, a picture appeared on the screen but it was weak and washed out with the same color issues I was having with the composite input. Problem I'm having with the color is in the areas of strong or saturated colors they're very bright and saturated on the screen but in areas not as saturated or more like a shading on say a scene with a wall or background of trees the color is missing or only shows the areas that "rise above" some arbitrary threshold like some amplifier is not biased correctly. Think of it like a contrast control is set too high and the gray details are all but lost. I've gone thru the chroma sections replaced resistors and screen bypass caps, all the voltages are correct compared to the schematic but I can't figure this one out. So what I'm fighting on the IF board is a low video level coming from the detector, it will not go above about 2.5V P-P and the AGC is very aggressive, I have almost no gain no matter how things are adjusted and any thing short of 100% on the AGC pot cuts everything off. Voltage on the IF tubes are within a few points of the schematic and the tubes themselves check good, tried known good subs and I get the same thing. Video/RF modulator is a Blonder Tongue AMCM 860D CATV head end that's padded down to -70 dbm. I'm going to do a raw video injection using my op-amp circuit again this evening right behind the detector diode and look at what's showing up on the grid of the 12BH7 video out tube just to cover this path and I will post my findings once I know more. |
Can you post a scan of the schematic so we can see what you're doing?
|
The best I can do is a hi-res photo of the chassis schematic, very little online on the Magnavox sets of this era, I don't know where my notes & schematic of the op-amp video interface are so I'm working from memory on this one, they're around here... somewhere.
Okay so this afternoon I went for broke and literally jammed the output lead of my video op-amp into pin-2 of the 6MU8 video out tube without making any other changes to the IF or the video detector side of the board, hit it with some color bars and the most "watchable" picture appeared on the screen. Going this way to me seemed like the most logical in terms of maintaining the DC loop thru the sound trap despite the impedance mismatch so maybe this is something you might want to consider with your Zenith. I couldn't find a free online schematic of your EC chassis so I'm no able to follow along, if you know of one point me in the direction and I will take a good look at it. So the op-amp circuit I'm working with is a simple non-inverting layout using an Analog Devices LT1252 in the 8-pin DIP package. Input is direct coupled and the gain is set to 8V P-P output which is then fed to a 100 ohm pot to adjust the final output level, I'm using the suggested layout of the datasheet and it's powered by a regulated and well filtered +/-9V onboard linear supply. The input signal is also directly coupled to an LM1881 8-pin DIP sync separator which can supply raw vertical, horizontal and composite sync plus a burst/back porch and odd/even output, early on I was using the v-sync to drive the vertical section directly and H-sync to the AFC diodes thru an NPN as an inverter. In this arrangement raster lock was absolute and the horizontal would remain locked thru the entire range of the control pot. The big downside to all of this was an uphill battle with DC restoration and this appears to be far more stable now using the brute force video injection to pin-2 (grid-1) of the 6MU8, schematic of the video amp below. Now I'm waiting on a fresh 6MU8 to arrive tomorrow, want to make all the adjustments using a known good tube. schematic of the video amp used in the Mag T-940. Flag 56 feeds the input of the 6GH8 sync sep and between R726 & R727 is the feed to the AGC grid-1. Plate section of the triode goes thru a parasitic suppressor then the delay line while the grid gets a blanking pulse from the vertical. Flag 22 goes thru a 22k resistor then to the low side of the brightness control which has the high side on a divider and +140V. For some reason imgur is not allowing me to post more images so I can't include the wider view. Okay an addendum to all of this. I have finally integrated my video/sync buffer amp into my Magnavox with excellent results, most of the work was housekeeping and finding the most suitable places to inject the signals and at 03:30CDT this morning everything came to life. I ended up connecting a shielded wire directly to Pin-2 of the 6MU8 1st & 2nd video amp tube leaving the entire video detector circuit intact and removing the 3rd IF coil connection from the plate of the 3rd video IF tube, from there it was run a pair of shielded wires from the buffer amp vertical pulse output to feed the vertical osc which was isolated from the sync separator by lifting one side of R-515. Horizontal is still keyed off the sync separator which has now become essentially an amplifier to feed the needed 35V P-P to the AFC diodes, one side of R-725 was lifted to inject the composite sync out of the buffer amp. The reason for this is twofold; one I can now use sources with non-standard sync levels to maintain a raster lock and two maintaining the sync separator maintains the original timing of the horizontal to the burst gate. I'm using the Texas Instruments LM-1881 video sync separator chip for the timing as they're very simple to use and have plenty of drive, the composite sync feeds a 2SC945 which is emitter coupled to the 6GH8 sync separator via C-724. The biggest problem I'm having is uploading images to the videokarma server and as a result they are not available for this post. If you want to see the images you will have to send me a private message with your email or other means of file transfer as downsizing them to fit the antiquated requirements cuts too much of the details out and is completely unacceptable in this day when the minimum image from a typical phone camera is several megabytes in size with little one can do to reduce this. The alternative is to follow along on another site where I have also posted my progress on this and many other projects over the years on my dedicated thread called "On The Bench" that is found in the Ham Amateur Radio section of smokstak.com and my username there is Brhatweed. |
A tube set might require as much as 4V P-P video.
I have to shrink my images down to 640x480 to get them to post. You can do this in "Paint". What OS are you using? I use a Canon 12Mp camera and transfer them to a Dell D630 laptop running W7. W7 has the software built in so you can transfer pictures from your camera to the laptop. I then crop or reduce them as required in Paint, then upload them via Firefox 32 bit to Videokarma. |
Running Linux on my end but regardless of the OS it comes down to principals. If videokarma wants support then they need to offer something to support. I resized the images down using Gwenview to fit the videokarma constraints shown on the upload panel and they were still refused, I'm not the one running the same software from before the karma split and have to say this is doing none of the members any service.
As for the set itself the new 6MU8 arrived this morning and it turned out to be a NOS dud that resulted in a washed out picture with retrace lines, going back to the original tube clears up the problem so it's definitely the replacement tube. Going to give it one more try and buy another 6MU8 as they're $5.95 while this one is left to burn in for the afternoon. The full write-up with large detailed images can be found here https://www.smokstak.com/forum/threa...5#post-2110545 Small images are in the post below Computer here is a HP EliteDesk G4 i5-8500 running ubuntu 24.04 |
6 Attachment(s)
Images for the above post. Tried uploading to Imgur but that's an even bigger pain so I've cropped them way down to a Windows 95 dial-up friendly size. If you want the real-deal email me and I will reply with the full size images.
My whole point to all this is to share the details and maybe spark some ideas on keeping the old sets going in the modern world of unfriendly tech. I'm still dealing with some color shading issues and will post them in the near future as they have me completely baffled. |
There is something wrong with your buffer detail. The emitter should be negative with respect to the collector if it is an NPN transistor. That circuit would call for a PNP transistor. You really need TWO stages of amplification. NOW, you can get amplification out of one stage but it will be INVERTED which is NO GOOD. SO, you really need two stages. The first inverts (with gain) and the second re-inverts. SO, now you are "right side up." Remember, in video, polarity matters.
|
That only handles the composite sync for the horizontal not video. In the bigger picture the buffer has been sidelined for the time in favor of the original IF strip as the op-amp itself doesn't have the swing the set needs on the 1st video amp. The issue with the IF strip turned out to be an open parasitic suppressor coil which was only letting a little signal thru on the internal 12K resistor and this was discovered using an IF sniffer connected to a later Panasonic TV.
As it turns out the 1st video is looking for a signal with 10 volts P-P to make the magic happen so an additional stage may be added at a later point, right now it's get a stable image on the screen and the chroma section working properly before I continue with the composite input. Problem I'm having with the chroma has to do with this green shading that seems to be some sort of 3.50 osc leakage that's getting amplified by the XZ demod and equally imposed on the R-Y and B-Y. It goes away if I advance the color killer or pull the 6HS8 demod otherwise it's always present and baffling me. Just another bug to iron out. |
The circuit I posted might be able to give you 10V P-P if you use 24V as your supply. Transistors like 2N2222 and 2N2907 are rated at 40V and should do the job.
What is wrong with the existing sync separator that you need an outboard one? The green shading, if it is not a purity setup problem, could be inadequate decoupling in a video circuit getting crosstalk from a sweep circuit due to a weak cap. |
Reason for the sync sep chip is so I can use non standard or poor sources and maintain a good lock, I also run everything on a master BB source and an external/internal sync switch was in the works.
Shading seems to defiantly be in the XZ demod section. I thought it was input video related but noticed on some scenes things appeared to have this sort of trailing color smear along with very intense reds and blues. The 6MK8 checks out and I have several new examples of this tube along with the 6HS8 demod. I can turn the color control all the way to zero and even short the pot to ground and I still get a shading that varies with the tint control. If I pull the 6HS8 or 6GH8 3.58 osc tube the shading goes away, where or how this is leaking in has me stumped. |
Defiantly, you say? That shading is definitely something coming across on your supply rail. Most likely your demod and 3.58 oscillator tubes share the same supply source. Scope it out. You most likely have a bunch of electrolytics that are weak in that set.
I want to see a scan of the schematic of the video detector and the first video stage. Can you post that? |
Quote:
I can post the schematic sure but will have to take hi-res pics to do it, would you like an email? |
You can try that. Those call letters at Gmail.Com would be good. But if its a very large file, it won't make it thru the email system. Can you convert it to a pdf? I think Linux can do that.
|
I will snap several detailed sections so you can see what's going on here. Give me a few days to get this all together. This morning was our official fishing season opener and a total bust for me. Cool & windy across Knife Lake.
Added some bypass caps to the DC feeds going to the various sections but the shading remains, going to put another day or two into this before I sideline things for a while as I'm getting burned out and need a break. I am starting to wonder if my add-on caps are bad/leaky/high ESR as I'm getting a very well defined 120hz hum, the hum bars are very square and distinct not rounded like a bad capacitor normally shows so later this evening I'm going to cliplead a couple of known good caps in and see where that goes. Never noticed this until the AGC issue was ironed out as it was masking things with its own issues. Set uses a full wave bridge not a doubler and a capacitor input choke so maybe there's some resonance with a failing cap, the original "cans" are out of the circuits and only plugging the holes and I'm using what was supposed to be low impedance Rubycon 450V photoflash caps (EC series) but they're now 20 years old having been installed the summer of 2006. I don't see more than a slight bit of ripple on the post choke side on the scope, manual shows 2V of ripple on the main 410V source to be acceptable so what I have is far below this. A possibility is heater/cathode leakage in V703 which is the AGC/sync sep 6GH8 so I will also sub this with another... I have over 50 new & used pulls to choose from. On the plus side of all this the IF sections appears to be on the money and the picture is about as good as I remember it minus the 120hz bending. Vertical is running a bit short despite swapping new tubes but overall I'm happy with the progress and may stay with the RF distribution using the CATV head-end Blonder Tongue modulator, its a CAMS-60 MTS modulator with an IF loopthru and I was able to use the IF from this to verify the Magnavox. Not exactly outlined in the service procedures but it works. I feed the modulator directly from a Microtime T100 TBC so everything is about as clean as an actual broadcast. |
1 Attachment(s)
Those were some very large files. Surprised they made it thru the email system.
Attached is a drawing of the connection I would use. You should go in right after the video diode. Disconnect the anode of the video detector diode and connect a 2.2 uf capacitor. Then connect your video source to that. Be sure to terminate into a 75 ohm resistor. That's a very old set. You probably have some other problems. I hope you have a good supply of tubes as they are drying up! |
I measure my tube inventory by the 25 gallon tote however my stock of 6LQ6's are getting thin. I was a Motorola radio tech for the better part of 25 years and accumulated a lot of RF service tubes along with the usual 6GH8s and related low signal TV tubes which go back to my youth messing with TV's.
The problem with the Mag is it's looking for greater than 8 volts P-P on the grid of the 6LM8 pentode, the schematic calls out a volt or so but so far experience says otherwise. I made an op-amp based input interface with 10db of gain using an LT1252 video op-amp using the mfr example schematic however it was hitting the rails at +/-12v driving the 1st video amp crushing the peaks. I may pick up the project again over the weekend if the replacement tuner shows up, USPS promised a friday delivery however that was pushed back to saturday. I used to be a rural route carrier and know low priority packages are not delivered on the weekends around here. The big picture plan is to get the replacement tuner in place, run thru the IF once more with the marker gen and bring everything back to "stock" as I've made so many mods over the 40+ years I can't remember them all. It's an original family set and I inherited it in early 85 when it was replaced in favor of a 27" Zenith S-3 however this one died from a shorted K/G1 short on the red gun and ran in this condition overnight. It had the A/V inputs and was left on with only black/burst and no one noticed the set was on, Magnavox won the longevity battle. |
Yes, V5a grid (pin2) is looking for a volt of video. Why do you think the schematic is not correct? Did you see that circuit that goes off to the Brightness control? I would investigate that. Check the plate and screen voltages of V5a. Verify the peaking coil is not open. That would screw up the bias on V5a.
Maybe a problem further down the road. How many tubes are in that set? Looks like a well designed set to me. I once worked at a radio shop. Did mostly installation into cop cars. At that time, it was Mocom70, Micor, and Maxar radios. Syntors had not yet come along. |
No the schematic doesn't reflect the same voltages as the T933 uses the 6LM8 for a video amp and my T940 uses a 6MU8. The DC bias from the brightness pot is present. My T940 print isn't nearly good enough for a photograph.
So the replacement tuner arrived late Saturday afternoon and installed without a problem but now I'm chasing down another that really has me scratching my head. Tuner has two B+ sources, #1 is the +130V source derived from R240 which is an 8200 ohm 1W on the supply divider board and this feeds the plate circuit of the 6HQ5 RF amp, the other is B+ #2 which is 140V supply that feeds the 6HB7 mixer/osc and gets dropped further to run the UHF tuner when selected. I'm finding +235 volts on the #1 B+ source and it seems to be from a lack of loading, the resistor itself checks good on the meter at 8225 ohms. I've swapped three known good 6HQ5 triodes and nothing changes, these are the ones that have the cathode tied to pins 2 & 7 via thin ribbon inside and are easily damaged by careless testing. I know everyone is passing a signal as I'm feeding in a known good IF signal into the UHF port of the VHF tuner and selecting UHF and if I feed from the UHF tuner I get plenty of mixer noise which is not present on any of the VHF channels. This is a known good tuner so I' at a loss with this. My test IF signal is from a 80's vintage Panasonic 11" color on an iso trans and what I see on this set is perfect, I tapped the IF out of this tuner and feeding a dedicated jack I use for testing. The original issue with the tuner was an accidental contact of the AC terminal of the little fuse block near the IF input with the tuner IF cable and that popped a number of parts in the tuner, thankfully another member was able to send me a spare replacement. The IF cable is good showing continuity and no shorts, it's passing the test IF to the chassis IF strip and all is well but it's like the RF osc is dead or way off frequency due to the elevated voltage which itself doesn't make sense. All the grounds are good and the AGC bias is right on the mark at 0.45v. According to the schematic the signal path in UHF starts at the little RCA jack on the side and is then coupled thru the RF amp thru the turret, from there it's coupled to the mixer (pentode side) of the 6HB7 which is acting as a second IF amp before it makes its way down the IF strip to the detector. IF AGC action is good and I can easily push that to an overload or cutoff so there's plenty of signal making its way thru and I know the Panasonic IF is spot on as they use a SAW filter to set the response and the magnavox matches this. I'm at a total loss to why the tuner B+ one isn't being loaded down by the RF amp. Gone thru three known good 6HQ5 triodes and no changes. |
Check for a broken pin in the socket or an open cathode circuit.
If it worked before, it can work again. An old CE once told me that. One time I pulled a fried rat out of his TT25BL transmitter. He came into the control room and looked at me and said, it's a dirty job, but somebody has to do it. Probably wasn't a good idea getting him around all that high voltage after he had just gotten back from the Gold Bug Room. |
| All times are GMT -5. The time now is 08:29 PM. |
Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2026, Jelsoft Enterprises Ltd.
©Copyright 2012 VideoKarma.org, All rights reserved.