sk97fet?

Overundr1

Overundr1

AK Subscriber
Subscriber
Another 717 on the bench, this one has a bad sk97fet input which under increased input gain goes full negative offset tripping the speaker protection circuit which was a bear to sort out in the first place. Anyways, that original fet is made of unobtanium so did some digging around. It appears that the current production lsk489 will be a drop in replacement with very similar idss values. Anyone else run into this who might have a better solution? The other route is to scan that happy auction place for a driver board (there are many that used this sony fet btw) to harvest one.
-Lee
 
Register to hide this ad
You are certain that it is the 2sk97? Have you swapped with the one in the other channel and it solves the problem?
It is a DC coupled amp, with LFreq AC/DC gain being the same, anything in that ckt could cause the problem you encountered.
It is a lousy design because they were to cheap or stupid to put in a DC servo, they did not want to use a ecap in the feedback path as they knew they could gain a small bit of LF THD performance.
You are sure those offset pots are solid and not misbehaving?

Good chance that a lsk489 or a lsk389 might to the trick if it is that but I would make 100% certain as I suggest above. I'd have to do some sims to know for sure.

Did you check out this thread
http://www.stereo.net.au/forums/top...ement-for-inputstage-of-tan7b-power-amplifier. Ended with no conclusion.

I know my G-7500 uses a 2sk129(L,M) and so does later G-5700,6700,7700

Now that I look further there is that crazy AC/DC coupling switch involved.
Try an external signal source that you know for sure is AC coupled.
What is the DC offset of the prea-amp, with and without the AC coupling in and out?
Lots of things to check.
Good luck
 
Thanks Rick
Findings----
The 717 has plug in connectors for the left and right amp cards from the connected/separated switch
No noticeable dc element on either connecter, ok maybe a mv or 2
Swapping input connectors makes no difference, fault stays on the f2722 board
Board completely rebuilt, glue removed, new trimmers, fault remains
Board will behave until input level reaches a certain point when the board goes crazy
Scoping the outputs during test shows no oscillation before fault occurs
Setting bias to zero delays the fault onset slightly however imo all this does is crutch the protection circuit
Loath to desolder and resolder those fet's due to possible static damage
Power supply dead stable, also measured dc over ac component across both the eq/phono board as well as the tone board and almost zero dc detected.
-Lee
 
I would just swap the FETs around to be sure, as suggested. (assuming everything else seems okay and you are unlikely to fry your 2nd FET).

BTW I have detailed measurements of the LSK389 and LSK489 going beyond what is in the spec sheets and will post more over in my thread here in coming days (or weeks...). But just beware they are pretty different units, make sure you order the right one.

There are different versions of each of them, which dictates Idss but Vp is also affected it seems.

For example:
LSK389B: Idss ~7.8mA (Vds=10V,Vgs=0V), Gm ~ 21 (Vds=10V,Vgs=0V), Vp ~ 0.62 V (Vds=10v)

where as for the LSK389C I measured bigger I_dss of 13 to 16 mA for 4 units (they also get hot and are quite temp dependent so I am not sure I like them - too hot to use at V_ds = 20 V for me. but they do have negative temp response (I_dss goes down as temp goes up) so they should not 'run away'). Vp was a bit bigger than on the 389B too... typically around 1.25V compared to 0.6V.

For the LSK489 I tested, Vp for those units were ranged from 2.1 to 2.5 volts. Idss was 5 to 8 mA.

Personally I would like to measure some sk97 myself before making sub recommendations and check Idss in particular. From what I see on the specs people provided on that stereonet post, I might be inclined to try LSK389B or C first, but really would measure a 2sk79 first and check. Hope the info helps. Also, would check the 2sk79 first as the definite culprit before doing anything.
 
Last edited:
Thanks to AK karma I now have a pair of sk129a's coming, will report on findings once I get one installed. Your thread on the lsk389 was interesting reading and they will be another source, in fact might order up a few just to have on hand.
 
Overundr1 said:
Loath to desolder and resolder those fet's due to possible static damage
Click to expand...
do you know a jfet was invented before a bjt was?
They are jfets not mosfets, thus are no more static sensitive than a bjt. But it certainly does not hurt to wrap some small #30awg around them to short and act as cooling to be 100% sure that you do not cook them. Old Ge bjt were very easily damaged by heat, treat like a polystyrene cap.
I am still not convinced that it is a 2sk129 problem until you change them. Anything in the ckt can go squirrely as the signal amplitude increases.
you can do as smurffer77 did and test 2sk129 out of ckt.
We all want to know what the Idss and Vp,gm is.

these are pretty simple tests to do. just have to be very careful. when G/S is shorted, the device is full on, lowest Rds and acts as a resistor. All you need is a R in series and a DC supply with an mA meter in series or measure the voltage across the series R and V across the Vds.
If Idss is 10mA, then you just have to make sure you do no put too much voltage across them. the larger the Vds the more power they will dissipate. we are talking 100mW max, they usually test them with pulses as not to heat them up.
So for 10mA Idss and a 1K R you need slightly over a 10V of supply V. carefully minotor the mA as you increase the DC supply.
You can determine Vp with you play with the bias V.
good luck
 
You must log in or register to reply here.
Back
Top Bottom