sx828 front end

[john stumpf]

in the alignment procedure for tracking, the first step is to check the b+ current drain 11ma +or - 4ma. does anyone know how to go about doing this? its pretty much unmolested and was used in my office playing music from my computer, the unit is dead quiet and aux and phono work great as well as the am radio but has suddenly developed horrible distortion in both channels with a lot of static but still has a stereo lite and a strong signal showing. there are 3 fet a 3sk35,2sk19 and another that isnt labeled on the front end schematic. there is no parts breakdown in the hifi eng, schematic. its rather hard to describe but instead of the usual 30+ channels it used to get its down to about 6 that are not coming in very well.so to my mind it must be a problem in the front end. i did a tracking alignment two years ago after a complete rebuild but omitted this step as i didnt know how to go about it im beginning to think that i might have a bad fet.of course i could also be totally wrong.

 

[rcs16]

check your antenna connections, try another tuner to verify that sig is good and it ain't that.
using a clamp on DC milli-ammeter or open the supply wire from the PS to RF and insert a milli-ammeter.
It is hard to say exactly what went wrong, there are 2 RF amps and a mixer in the RF front end. I assume the mixer is also a 3sk35 mosfet, but you could use a 3sk45 as a sub
I know it is hard to get into those RF FE's to measure anything.
You can measure at the fet sources & drains to determine if they are passing proper bias current.
var cap cleaning sometimes improves sensitivity, so does adjusting the trim caps. It is best to use a RF SG for that, as they say in the procedure.
Not sure what would cause an alignment to go bad all of a sudden.
As I usually say determine if it is a RF fE issue or a IF,detector issue. I have described the techniques in other threads and to you before.

 

[john stumpf]

so if i understand you correctly i am to install a milliameter in the front end b+ and read TOTAL current draw? i cant find any antenna problems,i use the coax connections,but made a simple loop antenna and there was no change save for reduced signal strength. it runs on the same line as my 1500td which works just fine. all the fet's are buried inside the v.cap and it appears the the whole business is soldered together via tabs on the bottom. none of this is any fun to access and i would guess if current drain is out of whack the sane thing to do would be to bite the bullet and and change all the fet's and the local oscillator while im at it.. the 949 project some months ago comes to mind i dont recall the the number oscillator which was subbed but i have extra's. the other fet's seem to be available on flea bay though. lets see what develope's.

 

[john stumpf]

mr fluke sez 26 ma. at 12.53 vdc applied. done as i understood it to be a milliameter in series between front end b+ and the source. looks like something has given up the ghost. assuming that the analysis is correct is a 3sk45 a better part as opposed to a a 3sk 35? both are available on fleabay.can the se3001 be replaced with a 2n5770?

 

[rcs16]

I was looking at the qx-900 tuner unit, same issues.
Sounds good to me, just check pin outs. but it could be the jfet as well, wakamolly, start with the first RF amp as it see the worlds interference.

 

[john stumpf]

well thank you very much! it would appear that some of the parts are made in china.do you think they might be ok? i managed to get it apart with a rather large soldering iron i use to fix radiators its a thread all in itself! but i did no damage. yep this is one job where you want to change everything if for no other reason than to not have to open it up again any time soon!

 

[rcs16]

Good for you, I do not know of anyone taking one apart before, would have to ask MTF if he knows or has done one before.
best is to bench test the unit before re installation. another challenge
at this point it could be anything, you really do not know if it is a RF FE or IF/detector issue do you? that is why I always recommend to treat each as a separate item and test separately. ie. inject 10.7MHz IF into the IF i/p and test that way as well.

 

[john stumpf]

what do you think about the chinese fet's? you've got more smarts than ill ever have i was thinking the drain current at least could be checked on the bench by powering it up with 12v battery through my meter prior to installing it. its just as well this happened as the shaft wipers on the v cap needed a deoxit treatment and even if you desolder the cover you really cant access them.

 

[rcs16]

I really can not comment on the Chinese fets, since i have no experience with them. what ones? point to the source.
good idea to power up on the bench with 12V and see if you can measure the S,G,D voltages to determine if they are conducting properly.
got a pic of one all apart?

 

[john stumpf]

there are 2sk19 GR's and BL's on fleabay as there is no parts breakdown in the manual which do you surmise is the most appropriate? working on the the pictures

 

[rcs16]

see what Idss grading is printed on the jfet.
you are going to fill in the blanks by looking at the parts on the assembly.

 

[john stumpf]

they dont have any letter codes on the 2sk19 so i went with the gr's.but it will take 3 weeks to get here. i have a spare font end which is hosed up as well. it was the original one out of this set. i powered it up and it draws 18ma, its better but not right. i know none of this is very scientific. but i guess the worst that can happen is im back to fleabay. im going to change those out one at a time and do a current draw test after each one before i install it. anyone know to post pictures on this forum?

 

[rcs16]

you can measure the DC static Idss of the jfet, just think of it as a resistor, so make a volatge divider and a low voltage supply, look at the Toshiba datasheet for the test circuit.
If Idds is 10mA, using a 10V supply put a 5K in series with drain, so that is +10V -> 5K ->drain -> g,s shorted -> ground return to supply, if it is 10mA Idss, Vdrain should be 1/2 of 10V = 5V
You can measure the V drop across the 5K and determine Idss of your sample. measuring RF performance, make up the test jig

did you try to change the 3sk30?
you can check if the LO is running, place it close to another radio's antenna(rx) or run a wire close the the LO, to see if it will quiet the rx, note the LO is running 10.7MHz higher, so fully mesh the var cap for the LO you are testing to bring it in the range of the rx, tune your rx to hear if you can get a quieting from the LO. if the LO is at ~88 +10.7MHz, tune to ~98.7Mhz.

 

[john stumpf]

do ypu mean 3sk35? no i wont have any part's for 3 weeks. there's not a lot around here to make things from or any place's to buy most things.i used my jeep battery to check the current draw on the extra front end! it was perfect it measured 12.58 volts (one cell has seen better days) my computer whiz is coming by tomorrow to get the pictures here.

 

[john stumpf]

well to those who are interested here you go.i must be getting old as bold as brass on the top of the 2sk19 is printed for all the world to see "GR"

 

[rcs16]

corrected 3sk35, you can use almost any dual gate RF mosfet as a sub. ones in the TO-72 make it easy.
thanks for the pics, will add to my library. made by Alps for Pioneer.
The gate 2 on the MOSFET acts as the AGC, max gain with gate2 at ~4V, gain lowers as the gate2 supply approaches 0V. some designs do not use AGC feature and tie the gate2 to ~4V, something to check and bias if you test the tuner on the bench.
You have the right solder tool for the removal job.
if you do change the mosfet, wrap some 30AWG soild Cu around the leads to short them out while soldering and out of circuit.
I have seen anything to go wrong, trimmers, contacts, transformers/coils open, resistors/caps out of spec, basically need to figure out what each gain/filter stage is doing. if you do an RF sweep of the FE you can see if the tuned stages are in synch and can be adjusted to center/align each one

 

[Watthour]

HOOOO-LY GEEBUS! A Weller/Lenk iron certainly would have plenty of heat and thermal mass to soften that whole mess of solder. It's just short of an acetylene torch...

 

[john stumpf]

if you ever take one apart you will understand why. and thats all i had anyway! and propane torches are OUT. i used solder wick to remove the solder on the outside but it still wouldnt budge finally GENTLE prying got it to move. after it was apart it was apparent that solder was also holding it from underneath. if this job pans out ill attempt to repair the retrograde. rcs16 your points are well taken. i plan on changing the lot.and check all the resistor's while im at it.with regards to the se3001.i have 2n5770 and 2n2369a on hand which is the better choice? the fet's i ordered are to-72 and are supposed to be nec items from a dealer in hong kong who also had the 2sk19gr.the first front end was toast (the one apart)so a year ago i bought one on ebay and that fixed it...... for a while. i almost bought another as there is one for sale now. but it seemed that this approach offered a more permanent solution. theres not much of any info on this 828 malady out there, but i cant be the only one. how many operating hour's does any semiconductor last? i have a peak dca55. is it in any way useful to try and use it on dual gate fet's, say one gate at a time on the removed parts? if i try to ohm out the coil's do you have any idea what sort of values i can expect?

 

[rcs16]

what a PITA, good for you taking the time to dissect the beast.
semi's life depends on many factors. we know RF mosfets and jfets can be damaged by static esp connected to the outside world. very few people are able to repair these tuner units as you are finding out.
use 2n5770 for the LO in place of se3001. I assume that you did not test the LO op as I suggested?
there is also SS9018 which about the same as 2n5770. 2sc535 can be used for the mixer as well,
NEC?, 3sk41 was used on the Sansui G7000 series. look at that FE schematic, they used a 2sc1047 for the Lo and mixer.
I thought that you had already worked on other RF front ends so may have some spare RF mosfets
you can measure the mosfet for static bias in a small jig, (follow anti-static procedures) follow the datasheet test ckts. basically need to independently bias each gate separately, look at the spec, they test Idss with gate1 grounded and gate 2 at around 4V.
A proper vendor will ship a RF mosfet with a shorting ring ( only can do so for a leaded version) or at least anti-static packaging.
As for testing coils, they are very low R, so basically a opens test will do. You need a RF bridge or test jig to test for inductance. it is those C trimmers that can cause issues. probably best to test them out in ckt, removing them could cause damage.

 

[john stumpf]

oh God no! you fella's are true rocket scientist's compared to me! i'm just a man between a rock and a hard place trying to do a repair on an item that otherwise has its only future in the bin! the local oscillator i never really suspected its just that as this thing isnt that pleasant to access i might as well change it as a precaution. i dont know how to do 10% of what you suggested. but every time you write something i at least have a vague idea as to what NOT to do. i didnt know that mosfet's are that sensitive to esd. im also beginning to understand why the whole thing is "canned" with more ground points than whiskers on my chin. nothing to do now but wait, win lose or draw i'll let you know how it plays out..i'm optimistic.