(Searched :) Power roll-off SSB mode on 858 Washington utilizing NTE235/236 Output Pair

[automator]

Hey gang,

I've got a power roll-off problem that's making me lean towards a potential bias issue, but I'll give you some background on this poor 858 chassis Washington.

Good:

Output of main regulator set to 13.8v
Recapped with quality electrolytics (this changed nothing in terms of operation/fails)
Mode switch doesn't apply 13.77v to required output section (temporarily bypassed until I can find a replacement mode switch)
Good RX/TX audio on AM
5W out on AM (no roll-off, even after 1 min dead key)
Good SSB RX

BAD:
Stuck at power rolloff on two-tone SSB test, or whistling/testing through mic.

So, the radio will key up and modulate just fine on SSB, but the power will quickly roll off to just a watt. I suspect there's a base bias issue , so I tossed a 180Ω resistor in to drop the 13.77v down, and was able to get the bias fairly close current-wise (within 15% of the updated figures). Still rolls off.

An interesting thing that is occurring is...the output transistors will get warm without any modulation in SSB. Then, if you test audio input, it will have skipped the dramatic roll-off and just put out 1W.

Any thoughts?

Thanks!

-phil
w6sfo

 

[SuperLid]

My first thought, is the NTE devices are crap.

 

[nomadradio]

The NTE devices started life as legitimate catalog parts with 2SC numbers on them. NTE's business model is to scrape those off, print their catalog number on it and list it in a cross-reference book the size of a medium-city phone book. Tricky part is what number did that part have when it was made? Decades ago, Sanyo, NEC, Matsushita and Mitsubishi all made CB final transistors. Only Mitsubishi stayed in it after the 1980s. Your parts may be bad, but if you did not set the bias trimpots, you may be seeing simple overheating of the parts' innards. Pull both bias-test jumpers, put your current meter in line with the driver side, then the final side. The trimpots should set for 60 mA more or less. Over the years the final and/or driver may have failed, shooting surge current into the input-side Base circuit. This can roach the bias trimpot and the bias diode, both. Collateral damage from previously-failed final or driver transistors can cause trouble when new final/driver parts are installed.

A good final and driver only work when the circuits around them are right.

73

 

[automator]

My first thought, is the NTE devices are crap.

Definitely crossed my mind! I was rather careful to remedy the unholiest of modifications and failed sections prior to installing these things, BUT...I did not test the new devices in a fixture prior to installation. It's quite possible the driver is working just fine and the final output is tango uniform. Thanks for the reply!

-phil
w6sfo

 

[automator]

The NTE devices started life as legitimate catalog parts with 2SC numbers on them. NTE's business model is to scrape those off, print their catalog number on it and list it in a cross-reference book the size of a medium-city phone book. Tricky part is what number did that part have when it was made? Decades ago, Sanyo, NEC, Matsushita and Mitsubishi all made CB final transistors. Only Mitsubishi stayed in it after the 1980s. Your parts may be bad, but if you did not set the bias trimpots, you may be seeing simple overheating of the parts' innards. Pull both bias-test jumpers, put your current meter in line with the driver side, then the final side. The trimpots should set for 60 mA more or less. Over the years the final and/or driver may have failed, shooting surge current into the input-side Base circuit. This can roach the bias trimpot and the bias diode, both. Collateral damage from previously-failed final or driver transistors can cause trouble when new final/driver parts are installed.

A good final and driver only work when the circuits around them are right.

73

Thanks for the reply, sir.

I agree that these devices really could be anything, and the notion that NOS items are all but gone along with the idea that purportedly NOS stuff is just chinesium driven to "beyond active mode" did not evade me.

I'd read on the document found on Scribd to set the base bias conservatively, so I did the 15mA/40mA set up as if these devices were the real McCoys. This took a bit of doing as the supply voltage needed to be dropped down, hence the series resistor in the supply.

I was hopeful to transmit again and see 4W AM, but then the SSB roll-off power thing took the wind out of my sails.

I'll keep at it or part it out. Luckily I have a few more NTE235/236 to throw at it. When those are exhausted, it'll be a parts rig.

73...

phil
w6sfo

 

[automator]

Fixed. After some serious triple-checking, I revisited my initial bias setting and found it was through the roof. There's a bad 300Ω bias adjustment pot on TR43. I hit the pot with some DeOxIt and it stopped freaking out for long enough to take a proper bias current measurement. I've ordered a replacement pot.
With the 40mA driver/15mA final bias adjustment I've got 4.5W AM, 13W SSB. Good enough for me.

Thanks for the replies...

 

[nomadradio]

Rock on! Gotta love it when a plan comes together. Congrats for sticking with it.

73