Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: [1] 2 Next   Go Down
  Print  
Author Topic: Correct biasing procedure and replacing MRF429 finals with MRF448 in an TS-950SD  (Read 4362 times)
DC1MC
Member

Posts: 13




Ignore
« on: October 15, 2017, 01:42:05 PM »

Dear Elmers, as I about to rebuild the final stage of a Kenwood TS-950SD (not X), I would like your "Sensei" advice about this operation, my background is in Electronics so you can be as technical as possible.
Beside the normal tools (milivoltmeter, power and SWR meter, dummy load 500W) I also have access to a Tek 200MHz - 1Gs/s with 4 channels, unfortunately no signal generator or spectrum analyzer yet (but a Siglent is in the Santa's list Wink.
So the Kenwood Service Manual is as cryptic as it could be, the schematic has on page 203 some DC values and the biasing procedure itself is somehow described in a table at page 162-163 in a way that I don't fully understand.

So to make the questions individually:
- do you have a description of the finals and pre-finals adjustment procedure, from start to end, beginning with the VR positions (probably to the ground, but who knos, as the DC biasing points are not specified, or I can't find them) ?

- is it any idle bias current (class AB) or all the transistors should be fully blocked (pure class B)?

- how could  use the scope to set the best wave form with least distortion, harmonics and so on (even on the expense of efficiency, because my QTH is in a densely populated area in DL and they are merciless with any kind of interference), any hints ?

- I want to replace the soon to be extinct MRF-429 finals with MRF-448, any special things that  need to care for, how much power/reliability/quality of signal is paying a premium for a MP will bring me ?

- If I replace the finals, should the pre-finals 2SC3133 be enough to drive 100W or do I need to replace them as well with something else, if so, with what ?

- Is the mod that installs the 6 x 1N4004 diodes in different places along biasing path worth doing ?

  Many thanks for your kind help to not burn some expensive silicon or produce some annoying ORM  Grin

  73 de Mircea YO9CZL
Logged
HB9PJT
Member

Posts: 355


WWW

Ignore
« Reply #1 on: October 15, 2017, 01:59:49 PM »

in case you would like some spare mrf429, let me know.

73, Peter - HB9PJT
Logged
DC1MC
Member

Posts: 13




Ignore
« Reply #2 on: October 15, 2017, 10:32:46 PM »

Hi Peter, yes, I'm interested to offer them a nice and warm home  Tongue, I've send you a PM.

Now back on topic, the 6 diode mod is described in this picture,


any idea if it makes sense ?

 73 Mircea
Logged
KC4ZGP
Member

Posts: 1637




Ignore
« Reply #3 on: October 16, 2017, 01:05:36 PM »

MRF 429 matched pair very available at RF Parts. Soon to be extinct?

I have an idea. Before you rip out the old transistors, read the bias voltage
found presently. Lift a base lead and read the bias voltage. Install the MRF448
and go from there. Don't over-think it.

Kraus

« Last Edit: October 16, 2017, 01:07:55 PM by KC4ZGP » Logged
ZS5WC
Member

Posts: 632


WWW

Ignore
« Reply #4 on: October 17, 2017, 01:51:06 AM »

 :)Make sure bias transistors are OK and bias pots are set to no bias, i.e. wipers are at 0V.(CCW according to manual)
Make sure 48V voltages are correct, other supply voltages are correct.
Bias for Pre-driver and drivers. Connect multimeter on mV selection over TP1 points.
Set VR1 to read 100mV accross TP1, which will equate to 1A bias for those.. (Sense resistor is 0.1 ohm)
Connect multimeter to TP2. Set VR4 to read 50mV  (equivalent to 0.5A) for one FET.
Now set VR5 to read 100mV on TP2.
The currents are additive--50mV for each FET.

BEFORE connecting new transistors, please ensure bias controls are working!..LESS than 0.6V at base pads.

73 de William
ZS4L / ZS5WC
Logged
KC4ZGP
Member

Posts: 1637




Ignore
« Reply #5 on: October 17, 2017, 12:21:41 PM »


I mean lift the base lead and read the bias current.

Kraus
Logged
KA5IPF
Member

Posts: 1406


WWW

Ignore
« Reply #6 on: October 17, 2017, 02:27:40 PM »

:)Make sure bias transistors are OK and bias pots are set to no bias, i.e. wipers are at 0V.(CCW according to manual)
Make sure 48V voltages are correct, other supply voltages are correct.
Bias for Pre-driver and drivers. Connect multimeter on mV selection over TP1 points.
Set VR1 to read 100mV accross TP1, which will equate to 1A bias for those.. (Sense resistor is 0.1 ohm)
Connect multimeter to TP2. Set VR4 to read 50mV  (equivalent to 0.5A) for one FET.
Now set VR5 to read 100mV on TP2.
The currents are additive--50mV for each FET.

BEFORE connecting new transistors, please ensure bias controls are working!..LESS than 0.6V at base pads.

73 de William
ZS4L / ZS5WC

That's fine but this is an SD, not a SDX.
Logged
ZS5WC
Member

Posts: 632


WWW

Ignore
« Reply #7 on: October 17, 2017, 10:13:06 PM »

 ;DYeah from the schematic above I see my mistake!..
Thanks for pointing it out Clif!..

Will have to re-think the SD procedure.

73 de William
ZS4L / ZS5WC
Logged
ZS5WC
Member

Posts: 632


WWW

Ignore
« Reply #8 on: October 17, 2017, 11:52:54 PM »

 ;)OK revised bias setting for TS-950SD:

Set final VCC according to service manual : VR3 set to 48V.
Make sure bias pots are set to 0v, CCW.
Set Rig to SEND, set TP1 with VR1 to 170mV. (Pre driver and Driver bias) equiv. 1.7A
Set TP2 with VR2 to 25 to 30mV. (250mA equivalent for finals bias)

73 de William
ZS4L / ZS5Wc
Logged
ZS5WC
Member

Posts: 632


WWW

Ignore
« Reply #9 on: October 22, 2017, 04:26:58 AM »

 :)Did you succeed Mircea?..
Logged
DC1MC
Member

Posts: 13




Ignore
« Reply #10 on: October 22, 2017, 11:15:55 AM »

Hello everybody, no, I din't go SK, on Tuesday 24 I have my German "prüfung" for class A, the highest ham class in DL.
The content of examination is light, but is in technical-academic-administration German, so now even if I'm reasonably 50+, I have again the old college dreams, trying to read an exam paper and not being able understand it, looking at the time passing fast and then I'm waking up sweating and start my HamRadioTrainer and other simulators and basically having the same reaction  Cry.

Reagarding the final of the TS-950, the Digikey parcel has arrived, two of the four BLW95 replacement for the finals arrived, the work table has been arranged, the board and chassis were thoroughly  cleaned and the revival operation will commence soon.

In the mean time, I've solved the 10ohm vs. 27ohm resistors, I'm putting it here for other people to find it: the TS-950SD seem to have had TWO versions of the final amplifier, the schematic of the low power version is given in the User Manual (4A fuse), with 10ohm B-E resistors and the 100W schematic is in the Service Manual (10A fuse) with the 27ohm resistors, the board ID is also different: the low power version is X45-3330-01 and the high power version is X45-3330-00. I have the high power version fortunately.

I have studied the schematic and understood the working conditions and roles of all the semiconductors and their biasing circuits (actually is pretty basic after you look on it), understood that the D1-3 are not simple diodes but 1.8-2.1V voltage references (practically 3 Si junctions in series), this why they were looking strange when measured as diodes.

Because I usually buy double and more for repairs, I may have some components left after  finishing the job (or not Wink, so I'll take requests from hams in EU zone if they need some original parts for reputable suppliers, list of available stuff will be published at the end.

But now I'll go back to the theory stuff, feeling again like student, that is actually a nice adrenalin feeling.

 73 Mircea
Logged
KA5IPF
Member

Posts: 1406


WWW

Ignore
« Reply #11 on: October 22, 2017, 06:26:38 PM »

FWIW the 950 SD only had one version (that I know of) and the power out was 150w. And yes I know the difference between the "S" and "SD".
Logged
DC1MC
Member

Posts: 13




Ignore
« Reply #12 on: October 23, 2017, 10:25:51 AM »

It seems that there is some kind of 10m only variant that has the PA a bit different and resistors with different values. May user manuals are floating on the net but the best that I could found as PDF was on the FreeBSD site.  Roll Eyes
Logged
KA5IPF
Member

Posts: 1406


WWW

Ignore
« Reply #13 on: October 24, 2017, 07:10:06 PM »

10m ONLY. Sounds like a US CB....
Logged
DC1MC
Member

Posts: 13




Ignore
« Reply #14 on: October 29, 2017, 03:42:39 AM »

Hello everybody, I thought it will be nice to tell you the end of my TS-950SD PA adventure, both to not leave the topic dangling and as a cautionary tale to others that will attempt something similar, so hare we go.
 WARNING: LONG

In the previous WE I got busy with the German HAM certification class A ( the highest class), that btw I've passed it and I'm now the proud owner of the cool DC1MC callsign  Grin. so it become very urgent to have some kind of working rig.
What follows is a story of agony and ecstasy and the treacherous and ruining ways of the sin of Sloth, AKA laziness.
So in the morning I cheerfully:
- cleaned the PA board,
- replaced the 0.1 ohm power resistors in ceramic package with nice Bourns metal ones in TO-220 package,
- in goes the new shiny 2SC3264 replacing the dead 2SC2922,
- one more full check of the semiconductors with an avometer (I believe the more experienced HAMs already see where this is heading), all (seem) OK, oh Loki you bastard,
- restore the thermal compound carefully and in minimal quantity,
- the PA board is screwed back into the chassis, more thermal putty around the temperature sensors, connectors reseated, compare the picture now and before, looks good,
- all the potentiometers (VR1-VR3) to maximum CCW as recommended,
- put a 10A slow fuse as in the Service Manual, connect the the Power-meter and dummy load (reports are that sometimes without a load there is a chance that the finals will auto-oscillate) 
- clear the area and press the power button...

SHINY VIOLET FLAMES INSIDE THE FINALS, STRANGE SURFACE DISCHARGES LIKE IN STARWARS MOVIES !!! (strangely enough, no smoke or smell)

FUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU....

AGONY !!! WHY, WHY, WHAT DID I DO WRONG ?!?!?

After a depression recovery break doing shopping with XYL ( on Tuesday there is a 2 day regional holiday here and everybody took Monday as a bridge day, the perks of living in DL Smiley I'm back in full Tony Montana mode: "You wanna f..me, huh, no I'll f...you",

- Out goes the PA board, damage control: finals - gone, AVR transistor gone, 0ohm across all terminals, no visible resistor burns, no other component measure as damaged (yeah sure, like I'll believe this s..t anymore).

- PA board goes back WITHOUT any finals, new AVR transistor, replacing the totally OK 10A slow fuse with a fast one,  power-on, no smoke or flashes,  input voltage 72V DC, output voltage with the VR3 fully CCW: 46V DC.
- set the output voltage exactly to 48.3V DC, connect a 470ohm/25W as load, rock stable, no drift.
- pre-finals bias, 165mV on TP1 seem to be prone to thermal runaway, WTF, is the Service Manual wrong ?!?!, what if my not REALL cheap DIGITAL avometer with 4K points it's actually a piece of crap for low voltages, to the rescue my old Italian analogue instrument spending retirement in a carton box, 1.5% accuracy, with 100mV and 300mV scale, the good Luigi measured 195mV what the digital piece of crap told me that is 165mV, set the bias to 165mV (equivalent to 1,65A at 15V, for both pre-finals, making around 800mA each, that is nicely biased class A).
- hold the key down for 1min, while checking the temperature and current, rock solid, barely any perceptible heating, good, now lets check the finals biasing.
- on the (empty pad) base - 8.2V DC !!!!, holy smokes Batman, wtf, where is this coming form ?!?!?, oh, the D3 2V1 stabilized junction that was connected across one of the finals is now floating, after a quick soldering to ground, the finals base voltage goes to 1.8V, OK, maybe that's because it's an emitter follower with floating output, let's put quickly a pair of 27ohm base resistors load as in the schematic, now goes to 1.2V, now we're talking, but hold on a second...
- I'M NOT PRESSING THE KEY DOWN, THE $HIT GIVES 1.2V ALL THE TIME !!! OH SHIT IT DOESN'T LET ME ADJUST IT !!! ROTATING CIRCLING DOUBLE §HIT, IF I PRESS THE KEY IT GOES TO 2.1V OVER THE 15OHM EQUIVALENT RESISTOR, Q7 IS BAD, NO WONDER THE FINALS HAVE BLOWN !!!
- RAGE, RAGE, RAGE !!! carefully remove the Q7, to not damage it, measure it again, it still looks like a transistor (2 junctions), then crashing it to pieces  while cursing 3-lingual obscenities, cat goes into hiding under the bed, is a curious but wise cat.
- OK, then lets replace it with uhmmmm... fark, didn't ordered any of these form Digikey, rummage trough the parts box, aha, a new KT805BM, so much better than the original treacherous 2SD1406, one problem though, the collector is not isolated, but the dead AVR transistor graciously donated his mica isolator foil and it's installed  Kiss, rest in peace you innocent one.
- Check the biasing, it now adjust properly with VR2, time to see what is on TP2 before soldering in place the finals.
- Put Luigi there on 3V scale ( a habit that saved my behind many times) turn on and CLAMP !!!, full scale, wtf now, 4.2V  on the current sensor !!! without anything connected ?!?!?
- Of course the 25W Bourne resistor that was supposed to be 0.1ohm is now 14ohm, oh well, back goes the old ceramic one that was saved, new measure something like 1mV without finals, good.
- Fix the fresh finals ( BLW95 ) on the chassis and carefully solder them using no residue flux, put the 220pF across C-E, the 27Ohm metal film resistors B-E close to the transistors, clean check, connect everything and POWER ON !!!
- Nothing, no extra voltage on TP2, no smoke, no output (that is OK, I removed the input to do the biasing in DC), big breathing session and here we go...
- Adjust the biasing current using VR2 to 25mV over TP2, equivalent to 250mA for both transistors, ca. 125mA each), no thermal rundown, nothing, all fine an dandy.
- Finalize the assembly, mount the shield, fold back the PA in position, connect the feeding cables and dummy load + the power meter.
- Press the key aaaaaand.. 5W out, huh, 5 W out ?!?!?, oh yes, I switched the power to 50W max and set the level to minimum, good then, back to 100W max and increase gradually to 80, 90, 100, 120, what the hell, let's not jinx it, it's enough, start keying: 5sec on / 5sec off to see the temp sensors, they are triggering, the fan speed increases,  all good, feeling a bit tired, WTF, is 3AM !!!, sleep time.
ECSTASY !!!

CONCLUSIONS:
- Laziness never pays, you always spend more time, money, efforts, whatever, if you do a thing, do it properly form the beginning.
- Check ALL the biasing and voltages BEFORE putting the expensive stuff in.
- Your fancy digital avometer may actually be crap at very low voltages and currents, this is why they still sell millivoltmeters.
- If some passives have been overloaded, like current sensing resistors, biasing resistors, whatever, mercilessly replace it, if you replace it, even if it doesn't look or measures damaged in circuit, the thermal and current shock has already damaged them.
- If you replace a part, always over-spec it, never under-spec, the manufacturer has spend a lot of time and money to make sure that the planned obsolescence kicks in after the warranty time.
- Have faith that you can do it if you work carefully and understand what you're doing, there is no magic involved, but of course if you have some black chickens you cam make a preliminary fried chicken Smiley.

 That's all folks, 73 de Mircea, DC1MC
Logged
Pages: [1] 2 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!