Cobra 148GTL This issue has me stumped

[Handy Andy]

Wow, I gotta' remember that someone else can come into a thread with a similar radio yet a different problem.
Wow.

My apologies for getting the two of you mixed up, Tolteca and Robalo - wasn't sure if both were the same person, just a different "sign on Nickname Avatar for work" versus one for home use, so now reading thru all of this I see we had two needing help - same radio - different reasons and approach.

Robalo - check the Cap's at the MB3756, when you shift off to USB or LSB - remember the loop frequency is much higher, like now 35 MHz or so - so that "drop off" may be a bandpass issue - where you need to tweak a cap.

So if you find the upper limit on the XCO is not making it - that may be due to the section needs to be more broadbanded. On earlier 1997 or earlier (Philippines Origin) radios had no problem going that high. The latter ones made in Malaysia are a little less broadband - tighter frequency band width - so the rolloff is steeper and if you have to - you may need to tweak L9 and the Pin 1 power supply for the 8719 - RCI models or the RCI 8719's needs a lot of drive to make the output peak out enough to make the external loop circuit fire and stay locked. So if you have an older 148 laying around or a schematic to it, note the parts values - especially in the loop sections for the UHIC 007 and the 8719's Timebase Xtal C85 and C86 - not only NPO types but the values may need to be changed to help the 24MHz Xtal send enough into the chip so it can resonate something in its output close to it.

Tolteca - Didn't forget you man...

Here's some more thoughts.

The 1.5kHz shift problem, lets' go back to that for a moment - did you see any work around the channel selector? The reasoning is this, if the detent isn't right - the PLL may think it's Pin programming is invalid and is stuck in one mode not realizing the shift did occur and it "Sees" a frequency it can't go to because it's being told it's out of lock or the Pin 6 logic level is marginal. So, check it to make sure the logic "floats" and is not being told to stay high - it needs the "ground" the channel selector has for it to "reset" internally it's own reference on that particular pin - hence "jiggle the handle - aka Channel Knob" reference. But, if that detent is dirty, it may hold the Pin 6 logic line at an unsteady level.

Also, are you aware that the year that radio was made, and if there was any mods ASIDE from the usual? I'm thinking you may not have a switch problem, but a failed cap problem - they (Dynascan) had several years where parts in the radio - did not meet the same on the schematic - for their reason of availability or whatever.

They put some radios' with bad discs, they show up "weird" in the DC-realm they're fine, but in AC, they're dead - as in shorted - and when tested with high-DC - they show breakdown short - even though they tested ok with battery powered DVM's. These were made of SR - semiconductor stuff...so they get punctured but test open like a regular cap. There are caps on the modes' power line they can change value quite a bit in their lifetime. I'm referring to C118 that electrically separates the oscillator side from the tuning side.

There are also some resistors of 10K values that were used like noise suppression on power feed lines (R237, R238, R239) that if old can "snap" and show open, this is an age issue, if they're AM'ers only that had owned this - it's easy enough to see that when the resistor blows open it changes how the other two lines can operate - so if works' been done to it or you have found several "blown" resistors of 1/6W - that may be another sign, that even in the power lines that are supposed to "knock down" voltage and isolate poles, dropping or pull-down resistors can fail.

Hey! When it's old - anything goes...even the diodes that power the TR30 L59/L23/L22 Mess can fail.

:+> Andy <+:

 

[Robalo]

GREAT now does not tx replace driver and final and tr41 same thing i think i goin to take a break

 

[Handy Andy]

Ok, you were taking it to the oddballs, just not going high enough.

GREAT now does not tx replace driver and final and tr41 same thing i think i goin to take a break

Don't know why TR41 would be a problem - but if it TX'd ok earlier, the issue of not TX'ing in the uppers towards the 10M band is a bandwidth problem. It's dealing with and around the PLL as a frequency issue, not a driver and final.

Ok, you are aware (right?) that there are more than one Xtal in this radio and you still haven't told me the year or origin, a board number does help, but only with the schematic and route, but NOT THE REVEISED PART LIST AND CHANGES MADE TO THE PARTS AS SUBSTITUTION AND UPGRADE...

The PLL uses X2 one series uses a 24MHz Xtal, - latter uses 10.240 - what year do you have...Knowing Origin does help me as Philippines or Malasia or other...because you are messing with upper limits - it could help in getting your PLL locked in...Hint NEC/TOSHIBA/RCI - all can use X2 as the 10.240 but not all can use the 24MHz one!

i have cobra 2000 gtl(PC-497AB crystal 11.1125) open voice lock -18kc +6 kc on the bench right now it has channel mods from 26.175 to 28.045 every thing is right on but only on upper 40/40 27.605 to 28.045 no rx/tx on am/usb and frequency counter drift .lsb is work ok.so i goin to check tr31 and tr32 plus diodes thank you Andy

When the PLL mixes it's own frequency with the IF (UHIC 007 side, not audio IF) - it looks at a window above 33MHz, and gets the image - then subtracts it using inversion (reversion is less applicable) or subtraction to restore the 27MHz signal - so you're trying to go way above, it's not in the cards the radio was designed for. It takes it up nearly past 36 MHz - there's an upper limit to this process before you lose the ability to keep the signal locked, and maintain some sense of a channel bandplan - for the multiplier ADDS to the error rate - lessens the accuracy of the PLL tracking the whole mess.

So you're seeing an upper limit on the frequency as to why the thing won't TX.

:+> Andy <+:

 

[Robalo]

date code 4 A 8 SERIAL 83000656 X2 10.240 (QX-07 88-B) I THINK FROM 1988 RADIO

 

[Handy Andy]

The PLL uses X2 one series uses a 24MHz Xtal, - latter uses 10.240 - what year do you have...Knowing Origin does help me as Philippines or Malasia or other...because you are messing with upper limits - it could help in getting your PLL locked in...Hint NEC/TOSHIBA/RCI - all can use X2 as the 10.240 but not all can use the 24MHz one!

When the PLL mixes it's own frequency with the IF (UHIC 007 side, not audio IF) - it looks at a window above 33MHz, and gets the image - then subtracts it using inversion (reversion is less applicable) or subtraction to restore the 27MHz signal - so you're trying to go way above, it's not in the cards the radio was designed for. It takes it up nearly past 36 MHz - there's an upper limit to this process before you lose the ability to keep the signal locked, and maintain some sense of a channel bandplan - for the multiplier ADDS to the error rate - lessens the accuracy of the PLL tracking the whole mess.

So you're seeing an upper limit on the frequency as to why the thing won't TX.

:+> Andy <+:

I seem to have missled ...

In the above post I'm referring to the Difference in X2 of 24MHz - not the Xtals' own timebase to 10.240MHz.

In my rush to get the post completed in-between tasks - I lost sight of the meaning in the paragraph...I'll take the ding for not proofreading this better...

So what I was referring to is the 24MHz difference between the 10.240 you're using and the required new frequency you have to apply to get the newer channels.

The VCO works on the 33MHz for 40 channels - you're WAY up - nearly 2MHz higher = pushing the upper limits of the frequency imposed on the VCO as well as all parts involved - to close to 37MHz if not higher.

So hence the 24MHz reference (~34MHz-10.240=24MHz) - even at 33MHz difference as triple the original timebase of 11.325 at X3 - it's evident that you are running into a rolloff / passband issue when the VCO is using frequencies that high.

Have you reworked the signal path caps the Tripler output uses? Mostly, the 47pF ones (C88 / C89) - else even the removal of a turn on L15 is the only thing I can think of to improve the higher bandpass the VCO needs to produce an output. The output is at the highest expected design of <36MHz - so you're working without a net...C88 and C89 - may need to be tweaked - as in a lower value, I suggest 33pF and re-check TR20 and TR29 for a better Hfe upgrade. The loop phase detect filter of C80 and C81 - are they Electrolytics now? Else the R100 6.8K might need to be lowered to as low as 3.3K in order to get enough drive thru the loop.

IT would appear that you got this thing to fire - just the losses even one of the SSB modes - means the drive of that particular mode is not as strong into the tripler.

IF you have a scope and a means to RF detect probe and run it into a meter - thinking the Noise Blanker section setup a dual diode doubler to increase the voltage detected and use a DVM as a mthod to maximize versus a scopes own problems with loading unless you have the tips for it - a test jig that lets you tweak the caps in the tripler section to improve it's output - else the only thing you have to work with is TP10 - not a bad way to go but it uses the TR20 output to show anything - I was looking further down into the tripler and into the LINEARITY of the two varactors that you are using on the radio - the two in that section I'm referring to D69 and D35 - both varactors are tricky, and one is padded with 18pF ACROSS as in paralleled with it affecting the overall performance of that varactor working with the tuning section one the goes to L59, L22 and L23.

So with the output - if it's not strong enough - is what you have done to it enough? I mean like - take a look at C116 - a 2pF coupling cap - that takes the entire 2nd varactors' output D69 out the Emitter leg across R144 (1K) that - and along with C212 acts as a tuning fork (peaking) for the base of TR30...sending it to the tripler...

I mean, be careful here - at that age - you're lucky to even get it to go into the 27.8MHz region as part of the channel plans.

It's why I asked about the age or year the radio was built - age is against you when it comes of the NPO and frequency sensitive parts, but it was designed back then using a hand drawn method that was very impressive in layout. The bane is it left a lot of areas with diminished capacity and ability to take the frequencies this board uses without inherent losses in trace thickness widths let alone the gaps between the trace routes

The above is for suggestions - not cast in stone - your mileage may vary...

:+> Andy <+:

 

[Robalo]

 

[Handy Andy]

This thread was about a 148...

I'm impressed - love the freq counter -

Ok, are you probing the DC realm - when you select the various modes, that's either the switch, or the power steering within the radio - you gotta' DVM those lines - that wafer is not perfect.

I'm curious if you've tried scoping' the line in DC to catch it using your triggered sweep - I'm thinking, now after seeing your vid - you have a bad switch. The trigger sweep should show it right away as a "jitter" or simple glitch.

I've seen this problem as a bad S403 - that is a ganged switch - and the logical portion of your problem points to the fact that a mechanical failure in the ganged switch for the 5th set of MODE contacts is most likely your problem. Why? Votlage drop - throw your scope on the line and try to see it's trigger. The 148 used a single layer wafer - the one you need is a dual. There is a SECOND wafer in there and if it's even warped - pow - no power or low power and you lost your AM. Since it's mechanical - there is also the issue of the physical alignment of the two wafers - one set not quite right and you have a resistor in line - not a contact.

Why its so important - the amount of power that particular line needs to sink thru it. Find the S403 and locate that 5th set of contacts locate the common and put a ammeter on it - I wonder in the MB3756 may have a rough time trying to keep the voltage up while that line is powering all those toggles.

Locate R152, check this "trigger" on both sides of this resistor - it's the one that goes to TR 31 - DVM it or whatever you need - if TR31 is having issues, you'll see it in this part. DC and Scope it to see if it "pulses" as it tries to come up - may indicate a failed part that transistor is fighting against while it tries to power up - heck even the R152 if it's as old as I think it is can drift in value.

Remember the R112 and R113 - the ones the route audio and bypass the AN612 - the 2000's show this as a 1.5K - I worry that when you see age in the switches - they tend to act like dead shorts and suck current away from where it needs to go.

It may not come as a surprise - but I wonder if someone "yanked" that knob in AM mode - damaging the wafer? It feels solid or does it have a but of play?

Wow - nice shop!

Ok, I could not see the 2000's freq counter when this Out-Of-Lock on the 7.8MHz occurred - I'm guessing you were on the standard 40 - so that PLL out of lock - should not have been "so wishy washy" there was one frame of your vid that showed 7799.9 - which was at 0 a moment before in AM mode. So, yes, that's a power feed issue - once the line and the switch contact in the wafer -stabilized - you got something - but only after a pretty hefty delay.



My apologies, I thought you were fixing a 148 - not a 2000 - but now that I see it - check S403 and make sure the wafers ganged together - are not too lose or sloppy as they may need alignment - this can cause that delay in the 7.8MHz uptick - TX otherwise seems ok as far as all the others.

Thanks for sharing this!
:+> Andy <+:

 

[Robalo]

when you measure voltage on IC1,IC2,IC3, ON WHAT CHANNEL YOU DO IT AND WHAT BAND AM,USB/LSB

 

[Robalo]

Back to the game

 

[Handy Andy]

I take it the switch passed?

You've done, replaced, a lot of parts - two finals a driver - several filter and audio amp caps and various jumpers...Diodes, even a 945C (Non-AQ though)

Ok, but how did the switch "Test" turn out?

The reason for the "Test" was that line AM mode, is the only one on the pole of the switch that has to effectively turn on or toggle three transistors, two varactor diodes, two coils and a re-route audio switch that would go into the AN612.

With the above in mind, what is the status of the switch? Is that S403-5th pole ok?

I guess I could ask this another way, how is the audio - does the audio sound ok in PA mode, I guess the SSB modes are ok, but what about the re-routing of audio In the AM mode - once it fires, and in either case, can you run an audio tracer to see how long, the delay or quality of audio passes thru that AM - Audio switch? (TR21) I'm trying to find out if there is any sort of problem with AM TX off of that switch - that is holding down the AM 7.8MHz for that long.

I am aware that you got LSB to work into the 28MHz band plan, but AM and USB won't fire up there...

So you know, unless there is another coil can you can swap, you may have reached an upper limit of frequency bandwidth the tripler the 2000 has - can give you. That's the LA256 can that's at 35MHz now...The DX models do this just fine (L14 but note they also DECLARE 10pF across the coil winds), but they "use" something that is made out of discretes and takes place of the UHIC007 IC2.

The only known ways I can see you even getting the tripler to push that hard, is to affect the bias and emitter resistors for drive in the tripler and even the output coil tank that makes up TP10...

The problem I see here is how the PLL loop and the tripler can mix without getting distorted artifacts - we can try to "Tweak" the circuits to make more, but when they get to their natural level or range of tune (their maximum resonance) it'll be so "loud" the output would be saturated - possibly a square wave. So linearity and bandwidth are what I'm trying to help you find. L19 - part of the UHIC - I can only guess is similar to the discrete design - but it's a 216 can part number - not a 165 - wondering if this is your roll-off issue...

Been looking over the 148DX models and they use a LA-166 - with a 10pF but what I thought was interesting is the use of the C66 and C65 caps off of the 10.240 xtal...they're 18pF - they use L15 as a LA-165 too...

So the PLL to fire that high. Can you locate R100 it's in that "phase loop" filter section - where the 2.2uF caps are for the detector pins C80 and C81, R100 is 6.8K while the DX uses 5.6K in it's MB8719 filter loop (R87).

I'm not bouncing these off just for fun, I'm doing some serious thinking over here in how you've run into a "wall" for the bandpass and that tripler. The DX's do this every day, but use 15 MHz xtals and a doubler to make this work.

So, I'm fighting two fires - one over the AM "power" issue, wondering if you have a sagging / dirty or otherwise troublesome pole of that mode selector. The other is with a bandpass filter issue (2-fronts of it) where the "skirt" you're trying to "Skirt-Around" is dealing more with the IC2 and L19 or are you have more issues with the Tripler?

Here's the rub, if the quality of voltage feed from the poles of the switch is poor enough, no matter what we try to compensate for in the output of the tripler, or even in the sensitivity of the PLL detector to make it lock - the bind is from the lack of power or not enough power - to make the tune-tanks/caps n' coil - design being sent to the 11.325 coils from any of the modes will hinder the efforts by lowering the output and affecting the output of the circuit as a whole. It may send a frequency, but it won't be as strong in output to make the stages amplify it enough to make the PLL see it.

So, I'm back to you - does S403 - pole #5 (or if it's a single ganged switch #4) - resistance check - is it consistent across all modes? Hence the scope reference would be a quick trigger sweep check on R152 - 2000's it's a 10K. Same of the USB mode - but at those triad of resistors/cap by L59 and VR6 (TX FREQ) - some of these radio chassis for the 148's don't even have them installed but if you can scope these or measure them BACK to the pole's common - and if they are uneven ohmic readings across the modes - we may have something.

SIDEBAR if only for a moment:
RCI / Galaxy uses the CPU to digitally "toggle" functions by routing a High-Low logic line to pass transistors - not unlike Darlingtons - to switch circuits on or off. IF you look carefully at some of these designs RCI used thru the years, they used diodes - varactors - on the switch power lines - to reduce noises from the switching, transients or any CPU hash noise present - for when they're powered across the feedline, they exhibit some considerable amount of capacitance that although buffered from the power supply feed thru a dropping - current limiting resistor - they provided some noise abatement but also can be used as a means to affect the tuning of circuits they're designed to turn on or off in-line with other RF passing circuits.
They were not meant to be used as Zeners - but wind up getting toasted when things went wrong.
The function the switched line does using this digital method is admirable - and takes away a lot of the guess work for analog based systems of earlier designs - but added a dimension of zany results and loop issues versus from the older analog high reactance element of the switching design.​

To step back a moment - Tolteca asked about a resistor reading on one type of board R158 - he's thinking a 3.9K while the 2000's used a 10K - note it's location and what it does.

Xtals are a funny thing - they absorb energy as a high-impedance EXTREMELY low current draw - but that doesn't help the tank circuit it's running - so the R158 should be a pretty high value - you don't want to lose power to obtain that 28MHz goal - so you may need to look around on the board to see the values used are correct to schematic and parts list - some service manuals are "not exact" so they use the disclaimer that "Circuit is not used in all versions" and of course the opposite idiom "Circuit Used In Some Versions".

SIGH...

:+> Andy <+:

 

[Robalo]

I have too many working hours on this radio, too many parts have been replaced, dont get me wrong but the radio TX better than my TS940 the audio is so great. It has some issues on RX sometimes it has a popcorn sound it has a lot of corrosion's and jumpers parts, you know what parts I've replace because you know the parts, I feel like a mechanic just replacing the damaged parts but not what is causing the damage. You asked me about s403 switch from 1990 cobra to 88 and same results. I replaced R100 (6.8k) for 3.3k in order to get enough drive through the loop. I also replaced c88/c89 from 47pf to 33pf and it was worse. Then i was looking for 100pf but i didnt have it so i found 68 and it is more stable. I also replaced the varactors d35/d69, also I replaced L19(LA216) and cut 1 turn of the ferite cord. IC1 pin 10 of the PLL and tie it to pin 17. I used diode 1N4148 inline, cathode (banded side) goes to pin 17. I think this loop is not working for some reason. also I think theres a bypass cap on short somewhere. Also FT-1,R26/ R65/R66/R67/R69/R70 are on bad shape.

thanks for your help and tips! greatly appreciated, I read each paragraph 20+ times! good learning experience

 

[Handy Andy]

Ok, with what you said over the "Xtal swap" then the bandwidth issue is somewhere else - it peaks out on normal channels, but it looks like you've run into the upper limits of bandwidth I can help you with.

So the C88 C89 by TP10 didn't help - so that wasn't the problem - thanks for getting back with me on that. Now I know...

Ok, then you're in a realm of bandwidth "drive" that you need to make choices over. You've achieved nearly 2MHz, but without serious rework, you are done pushing - you've lost a lot of parts over this. You can successfully say that you did up just about every channel you can get, in-between the 11 and 10 M bands.

If you get time, search for the Cobra 148DX - it's a Tri-Bander - nothing big, but the IC 2 UHIC is made up with, broken down into, discretes. PC-879 board. It's not necessarily the circuits I'm suggesting, it's the coil numbers used - Cobra had to limit bandwidth on purpose, so they went with cans with tighter (read : Less Bandwidth) "Q" so it wasn't even Cobra that had the issue, the supplier - Uniden - did. Which would explain the "tight lipped" stance they've had since then. Perhaps even as Lobbyist's for RF Spectrum...hmmm

Galaxy does up a similar design for the 145106 chip - EPF000210Z - not a pin for pin but with them doing it on their own, kinda' categorized their slant on keeping things going when the main producer stopped production. Look for L14, their version of the VCO "peak" coil.

You've swapped out about everything you can - and then some so - back off - take a breath and look at bandwidth - see if you can locate a VCO coil or two of the size outline and winds you need to finally push the upper limits.

For 2nd harmonics, a 16MHz VCO gets you 32MHz but a 17MHz gets you to 34MHz - best results, if possible, with a 18MHz (if you can find them - or wind your own) - but it's a long shot. Again, you're pushing upper, at the expense of the lower end - so once the swap, you may not get the full 11M band back.

Sorry to see you get snagged by this bandwidth - I know you can do this. But if I could make my own boards and have the machines to do it - I'd be up against some of the more prolific manufacturers like Uniden - they seem to have turnaounds like overnight - BAM! something new in the GMRS or Marine - CB is put off and thrown under the bus...like an afterthought.

Oh well - good luck, have fun!
:+> Andy <+:

 

[Robalo]

TR21,R112 R113, (AM) SSB D6 AND D7 ///////Handy Andy ////Otherwise TR10 and TR11 are "on" all the time fighting your efforts...and upset the bridge that helps with the "deconversion" of the SSB stuff...it really screws up AM reception but, if you don't see these colors on your TV set... I THING I GOT IT . GO ROBALO LOL

 

[Robalo]

This is how L20 and L21 reacts on 27.605 to 28.045. So tr20 which is the loop mixer is not stable. I replaced r101 for 9.1k. It seems to work a little bit better, any tips?

 

[Handy Andy]

I like the way you think - for now we're affecting the "Q" of the power choke - 9.1K and you're seeing something...

Ok, thanks for staying with me on this - for the ratio of R101 (original a 10K) and R138 (1K) - by L21 - are factors of 10. You just changed the Collector side...you are affecting the amplification but this will come at a price.

Another spot in an earlier message I talked to you about...



Swap R101 to a 15K to see if it "sharpens" that TP10 result. No guarantees. Just wondering if Current Starving it would help "smear" the waveform so PLL can see it. Again, we're going into 2MHz - usually this thing goes to about 1.8MHz "sample" bandwidth as its' upper limit.

And for tolerance issues, pull and measure R138 by L21 - you may need to install a fresh one here but ONLY after you measure - because the "window" and drive level are going to change with the fresh part. So you know, you are affecting an impedance level output - raising the ground resistor changes the P 2 P levels - but it comes at the price of Base Bias drive and linearity not just the waveforms power levels but voltage "Window" the thing works in.

Work slow, take it easy - nothing too fast - take your time.

:+> Andy <+: