The loudest AM "CB" chassis ever? EPT3000-10A

[loosecannon]

Wow Andy, now we have something to chew on!

ok, starting with the heatsink and the grounding of the tabs of the 2003's; i did find it strange when i didn't find a ceramic or plastic shoulder washer around the screws that mount the 2003's to the heatsink. There was no mica insulator, but after not finding the shoulder washer, i wasn't expecting to find them.

the screws that mount the 2003's are course threaded and thread right into the aluminum heatsink. This, and the pinout of the 2003 led me to believe that the tabs were indeed intentionally grounded to PC board ground.

If you are saying that they should be insulated, then i will try it, but i will at least state again for the record that the radio looks to be designed to have them grounded to PCB ground.

I will measure the idle current coming into the radio tonight when i get home.

I also have no problem with tying the PCB and chassis grounds together if that eliminates any of the feedback paths.

It looks to me like you want me to change R192 to more closely match the value of R190, is that right?

if so, (i am certainly not challenging your idea, but just looking for more insight) why did they choose values that had an approximate 1:2 relationship?

we also want to half the value of R191. should one of these steps be taken and evaluated before making the other change?

I would also appreciate any specific tests that you would like to see done in order to quantify the changes we are making?


I still have to get the PLL working, but im getting closer. it is a tough road for me though, since i dont have a lot of reference material to draw from.
LC

 

[Handy Andy]

Ok, in for a second breath on this...

Caught your message while I was posting this one...
I'll address the "tab to heatsink issue"

Not sure where to take this - but if you didn't find any - my research led me to places where they used the finned heatsinks but all had insulator and washer to tabs to keep the ground to foil grounds. This was more of a noise and whine issue but you did mention heat, which had me concerned that the Pin 3 ground and Tab grounds - tying together electrically at tab and at foil - a loop of noise or an unwanted feedback path - brining in whine, noise and possibly squeal issues.

The differences in values in both feedback routes of the amps R190 and R192...

This dealt with the efforts of Q4 in both - remember the R7 is the amps own internal feedback resistor. So to "Bridge" these things - the external cap is used to exert a reactive frequency-based (a working impedance on effort for tonal response) on one amp to INVERT or Push while the other Pulls thru the load. If we go too high - one of the amps will "smash" it's signal into the load and possibly generate a phase and potentially damaging power loop that is not unlike the TWIN-2312's and how the bias - if not done right - reduced power and shortened the life of those Finals too.

I think most of this problem deals with the T1 design more than the amps themselves. We have to make the bridge more of a balance between the two amps - and part of that problem deals with the voltage regulation -and why that little circuit - which is used in Galaxy's today - works like it does. The thing regulates when the SOURCE power is higher than the expected output of the regulator - once the SOURCE power drops to or at or below that set voltage - the thing no longer regulates - but acts like a pass-thru design only the diode voltage drop hinders it - no wattage losses. Other regulators would shut down - this one doesn't.

So if you see "boxcars" we need to reduce the gain of the amp - both of them - I'd do it equally and see if that can generate some symmetry so at least we would have something predictable.

Just my thoughts on your message - keep up the good work and thank you for following through on this...I want to see this thing work better than it did originally - if not the same as. I wish you all the best on it...

Continuing on...read thru then decide - the rest of this is the "why" I said the above...

I want I need someone to work thru this with us. The Issue I see is around the "AMP" itself - not the external circuit - the INTERNAL one.

Note location of Pin 1 to base of Q4 -skip the parts as needed but the leads of Q4 are what I'm concerned with.

Note R7 - goes all the way back to the output tie off of Pin 4 INTERNAL feedback path. Means this thing will have a preset "set" already - we are just providing an signal EQUALIZATION path - tone, Treble Bass - everything in-between. So again we are setting a gain and EQ affects that gain.

R 7 is already at 1/2 supply (on one side of it Vdd other Vgg or Vss) - as well as whatever is present in output on Pin 4.

Pin 2 AFFECTS what R7 carries into the differential amp Q4 - one side is tied to a gain source - caps and resistors to provide a working effort or gain - the Base accepts Pin 1 biased internally and forward biased as well.

Pin 2 though - is the one that will set how much Q4 will either source or drop across itself in signal - onto the rest of the amp. Q6 andQ7 are a darlington paring and Q10, Q12 are another for each share 1/2 power of the supply rail provided at pin 4.

Pin 2 Ohmic reference - refers to Ground, measured as from Pin 2 to Ground.

The problem is how much gain do we need in this amp - to make it equal yet balanced?

If we set Pin 2's ohmic reference too low, too much signal will arrive to Q1 base and set to Q6 and Q7 pretty close to maximum. (Your boxcars)

If we set Pin 2 too high, R7 "swamps" out any difference signal and Q6 and Q7 are driven on but at heavy idle current running thru it and low signal output. Heavy feedback loop correction - idle current. (sitting pretty but baking)

If pin 2s' ohmic reference is too high and not enough differential or opposing signal arrives on Pin 2 - Q4 sends the DIFFRENCE between get to Q6 and Q7 making the thing run hot - but too little and the thing runs a lot of idle current thru itself.

Perhaps a way to look at how COBRA 29 LX/LE's do it...

​

So, to keep that at bay, do we -
A. Low ohmic Pin 2 (like R501 above - which PULLS DOWN R7 signal level not voltage divider supply)
B. High ohmic Pin 2 (like on the SS33 16 ohms - pulls on R7 but not as much on Signal - nor Voltage divider)
C. Lets' all have a Beer and call it a day...As Seen On TV.

Ok, some will have to just ignore Option C...

What makes the above different than the test circuit - is the R500 and R501 ratio - pretty high and a lot of gain can be obtained - but there will be consequences.

One - the output goes into a highly inductive load.
Two - That inductive load will have a reactance towards any high-frequency components.
Three - IF not compensated for the impedance of the load will cause the amp to heat up due to reactive components based upon frequency response the load will exert as impedance (bass to treble limitations) and damage the amp

So they placed EQ filter in there to offset high - frequency issues by applying this filter as a hi-cut - anything above 1kHz or so is attenuated - starts to drop off. Q4's differential pin with R7 on it never sees any true low-frequency - nothing to invert - all high frequency (alto, tenor range and above are present) are put on Q4's oppose pin and with R7 - attenuated.

So they pretty much let the mic amp amplify high tones, and the Audio amp amplifies the low tones, and what we hear is the stuff that didn't get rolled off above or below - our bandpass.

So perhaps the better options are...

Increase capacitance values of the feedback gain loop to influence more on R7 (of each amp)?
Reduce Resistance gain values - the output of Pin 4 - say 220 and 2.2 ohms may be a bit low, but we are working on Pin 2 ACROSS Q4 and R7 internal resistance. - so push ohmic results to match the OHMIC WINDING measurement of T1's FIRST WIND on both amps?
Adjust Pin 4 resistor (at pin 4 output) as needed to reduce boxcars and over-amplification - keeping both equal until a balance between heat produced and audio gain needed?

Just some thoughts...
:+> Andy <+:

 

[loosecannon]

well your thoughts have given me a lot to think about.

and i will take option C but probably keep working anyway.

let's see if i can simplify some of your ideas so that i can understand what a am taking on at a given point, then maybe i can try one at a time and post the results.

first thing i wanted to make sure i pointed out was that i have not aligned anything yet.
the AMC pots for both high and low power are probably cranked up causing the box cars on the scope.

so, the first thing i will be doing is setting both for 100% modulation and see how that affects temperature.

second, i will measure the resistance of the first coil in T1 in order to give us a place to start with component value changes.

once i give you that value, im guessing it might lead you to suggest a couple different resistor values, so we'll look at that first.

next, i would like to play with the value of R191, trying a 10 ohm resistor there first.

let's stop there and see what sort of data i can provide based on the changes made so far.

what would be helpful for me is if you could help me understand exactly what to look for with each change.

right now it seems like running cooler is the goal we are after so i will establish a temperature rise rate of sorts and see if we can't improve that.

i will only have my multimeter temp sensor lead (will have to find it first!) to use, but at least it will give us a temp rise over time basis to start from.

one last question.

i want to make sure i understood you when you were referring to the bandwidth of audio freqs that this bridged amp setup is expected to handle.

are you saying that if i were to increase the bandwidth, that the amp might heat up faster due to the wider freq range it has to deal with?

thanks again.
LC

 

[Chevboy0167]

On a side note... found this. A few add on's in a kit. Double modulator! LOL

https://www.ebay.com/itm/Modificati...p-Colt-Double-Modulator-Colt-120/282803033830

 

[Handy Andy]

LC - in fewer words, YES - you widen the bandwidth, it may make the radio sound beautiful - but it's going, working into - an inductive load - a lot like AM only radios like Cobras and Unidens low-end series.

It the chip will heat up and even melt the lead solder from the heat the chip transfers to the leads.

So that is why I showed the Cobra 29 LX/LE series using the TDA chip - they have TWO paths for feedback and gain control - which makes the INTERNAL R7 value a lot lower - less gain.

In the SS33 you have there is only one and there are two values and their response curve involved. So one set of R/C values provide a tonal "Sway" while the other is close but has more gain for the "Sway" because R/C is different.

Best to stick with what was known to work at last time it worked and was stable - we can add or reduce more later.

Again, bandwidth is the heat problem too much and she will overheat.

And thanks in advance for measuring T1's ohmic - only hope that it can be matched easily enough and if my guess of 16 ohms is right - they used that ohmic measurement as their baseline for reference to ground. IF it's not correct - we use what we measure - then adjust the R's from Pin 4 (R1) into this divider (R1 and R2 into Cx) to reduce gain but obtain stability.

Cobra AM only ones that use a T1 setup - also use a R2 to ground and it's low but that is from the expected ohmic winds for that particular "tap" not the ohmic result of the whole thing...because the "whole thing" never goes to ground - it floats at T1 input from Amp and Source power supply.

This has been a fun learning experience and teaching moment...

 

[Shadetree Mechanic]

LC - in fewer words, YES - you widen the bandwidth, it may make the radio sound beautiful - but it's going, working into - an inductive load - a lot like AM only radios like Cobras and Unidens low-end series.

It the chip will heat up and even melt the lead solder from the heat the chip transfers to the leads.

So that is why I showed the Cobra 29 LX/LE series using the TDA chip - they have TWO paths for feedback and gain control - which makes the INTERNAL R7 value a lot lower - less gain.

In the SS33 you have there is only one and there are two values and their response curve involved. So one set of R/C values provide a tonal "Sway" while the other is close but has more gain for the "Sway" because R/C is different.

Best to stick with what was known to work at last time it worked and was stable - we can add or reduce more later.

Again, bandwidth is the heat problem too much and she will overheat.

And thanks in advance for measuring T1's ohmic - only hope that it can be matched easily enough and if my guess of 16 ohms is right - they used that ohmic measurement as their baseline for reference to ground. IF it's not correct - we use what we measure - then adjust the R's from Pin 4 (R1) into this divider (R1 and R2 into Cx) to reduce gain but obtain stability.

Cobra AM only ones that use a T1 setup - also use a R2 to ground and it's low but that is from the expected ohmic winds for that particular "tap" not the ohmic result of the whole thing...because the "whole thing" never goes to ground - it floats at T1 input from Amp and Source power supply.

This has been a fun learning experience and teaching moment...

Hey guys, this is a fun learning moment on my end. I don't completely understand all of it but I get the jist of it. I don't know much about amplifiers but I do remember OP amps in school. lol

 

[Handy Andy]

I'm having a hunch there is a peculiar reason for the use of ohmic values for gain and their differences between the two amps.

One drives "envelope" for the other.

If any of you can take a moment to look at any AM regulator in any SSB radio - the Cap resistor combo - is your gain cell as I call it. It stores energy for the AUDIO - but the energy itself is for the ENVLOPE - a place to put the audio signal into (envelope) is a slower - longer wave than the audio itself.

In a way - it sets upper and lower limits on how much audio drive is present in the signal.

It can be thought of a a "range of scale" - one amp is audio frequency all the frequencies and power, while the other side works as a lower end or cutoff (near Bass side of audio scale) and the upper limit in frequency is the inductive event in the Transformer (Reactance modulator) - so the audio fills in the "gaps" left by the lower Bass response filter that the other part of the amp provides - a sinking current with a bandwidth product.

The upper end is trimed off by the inductance (a reactive event - impedance - a hi-end hi-cut - rolloff) while the lower end works on the amp side that is NOT DRIVEN directly by audio input - it ghosts the other amp at a lower pace, or frequency response - leaving room for the audio amp to push in a range or bandwidth of frequencies the ghosting side.

You can see where I get this idea by looking at the values of the R/C networks - the "ghost" side is driven purely by a response to Pin 4 thru R7 internal from the other amp ACROSS from T1 - and the "ghost amp" external EQ feedback loop - it's not the same frequency response - - it peaks at a lower frequency range - towards Bass. IT is also higher in gain - leaving some headroom to drive - out of phase - a envelope that can be thought of as a pocket for high frequency audio to be placed in.

I found a lot of data here...
http://www.circuitbasics.com/complete-guide-to-build-a-10w-amplifier-with-the-tda2003/

Take your pick...
:+> Andy <+:

 

[loosecannon]

Finally had a chance to measure the resistance of the modulation transformer last night.

I have a paying job on the bench and my scope just bit the dust so this radio is slowly finding its way back on to the shelf.
I am doing my best to keep it off of there!

anyway, i removed the transformer from the chassis, just to make sure i didn't have any unwanted PC board connections being made after desoldering.

i measured each of the coils in it (3), and they all ohmed out the same.

each coil measured 0.2 ohms.

that is all i had time to do.

Andy, i figured i would get you this date first, and see if it changed anything in the order of operations i laid out earlier.

otherwise, i will just continue on and change the parts we talked about 1 by 1 while measuring the temperature of the heat sink.

LC

 

[Handy Andy]

Ok, in other words they are not much to offer any ohmic resistance so this means T1 is more of a distribution transformer than a modulator - so the inductance part of this problem is a little different.

Did some thinking too - so if we kept the ohmic values of all the resistors in the amp the same, and all you did was change capacitance - that is why I get a little nervous about the self-oscillation issue - you changed the caps to a higher value - which means more signal input - which leads me back to those times when kids used to take out the 470uF cap at the Audio amp out to the Transformer and put in some ungodly large value cap and wonder why it didn't work.

How about we get the thing back to stock and working right and then see about generating a quantum field effect Flux-capacitor design for it...?

:+> Andy <+:

 

[loosecannon]

yep you got it.

(i kind of figured you would eventually advise me to do this, but i didn't want to change anything out of order)

so, stock 1000uF caps will go in, but not before doing a 1 minute temperature test on the heatsink as it is now in order to establish a base line.

i will get this done tonight hopefully, but won't be until very late.

I do hope that going back to stock will cool it down enough, but the off color coating in this area on the solder side of the board tells me that it has probably always run hot.

LC

 

[Handy Andy]

Well, I could see some thermals, but not like roasting marshmallows...

What kind of repairs had to be done?

 

[loosecannon]

so far looks like only PLL.

only turned it on for a minute before recapping.

radio still acts the same as it did before. will go into TX, but only transmits on one freq.

honestly havent done any testing on PLL yet because it will transmit which allows me to play with the fun part of this radio.

LC

 

[Handy Andy]

Yes, I've had toys like that too...LOL!

"Hey MOM! We got the car to start, but can you make it go?"

Ahhh...<Chuckle>

 

[loosecannon]

well with the paying gig off the bench, and the new scope repaired, i finally had some time to get back to this radio.

I decided to concentrate on the PLL section and get the radio up and running and aligned before messing with and testing the TX audio amplifier.

I was having A LOT of trouble figuring out how all the switching worked because i was SURE that the cut trace on the band switch PC board was something that was done by a user, and not by the manufacturer/ distributor.

Well i was dead wrong about that LOL. there was no way this band switch was going to work properly without that trace cut, and once i reconciled myself to that, it all became easy and the wiring is basically the same as an old Galaxy 33DX.

first i was able to get the band switch working, but when i would key up, the radio would go into TX mode as indicated by the LED on the front, but the VCO would start bouncing around.

That's when i realized i had to tack solder a power, ground, and 'wire 13' to the other side of the band switch. once i had that done the radio came to life and was operating on all channels and bands.

During all this, obviously i was letting the radio get warm, so it had been on for over 45 minutes when i went to peak the first of the cans in the PLL section.

As i laid my hand down to put the alignment tool in the can, the side of my hand briefly touched the heatsink around the two TDA2003's that are used for the TX audio.

IT WAS BURNING HOT!

i had not keyed this radio at all, so these things were either oscillating somehow or one of them is shorted.

the reason i say that one of them is shorted is that i was checking things out and listening to the TX audio from this radio on another.

the audio was there but sounded hollow and weak. i remember being able to achieve full power output from this radio (22 watts +) so at this moment i am guessing that just one of the amps has a problem.

that being said, i would replace both.

absent any service data regarding the voltages i should see at each pin, i don't really know how i want to proceed yet.
I am open to any suggestions.


i did notice one other thing that was odd, but again, since i don't have any service info, i don't know.
I couldn't seem to adjust the 16.2700 mhz signal properly.
i was trying to use L13.
it would vary the freq, but as you turned further it would start moving in the opposite direction.

can anyone look at this schematic and tell me if this is the right can to adjust for this freq?
http://www.cbtricks.com/radios/rci/ss_33/graphics/ranger_ss33_sch.pdf

thanks for any and all input.
I am off to start learning about the TDA2003's in order to see if i can test one for short.
LC

 

[Crossbow]

Good to see your getting this radio back in working order! It was a weird one no doubt.