The 170/74 in question was given to me by another DIY´er as he had matched pairs (B-grade) Not sure the exact origin of these but I do not believe they are anything other than genuine. TBT...
I will pull out the SK170 in question, test it to confirm short circuit.
Replace R3 obviously. Given the outcome I had (blown R3 possibly due to a faulty 170) what else could this snowball into?
I had 3 DMM in place, monitoring R7, R8 and offset.
The system had given readings and output prior to letting the smoke out and the probes were stationary while adjusting P1/2, they were only moved during power-off.
Yes, but: He has not found the root cause off the first burnout, Yet.
Check the values of ALL resistors, you may have swopped R15, R19, R17 or there is a soldering short somwhere. Find the root cause !
Then check the outputs and the protection transistors again. Orientation of the potmeters P1, P2.
Basicaly you have to dobbelcheck the intire pcb-board before slowly applying supply voltage again.
Correct, I haven't yet.
I will be checking everything before firing it up again.
What I have done so far is remove R3, R2 and Q1. R2 is next to R3 and appeared burnt so I removed it as a precaution. Both R2 and R3 measured to spec. Both R2 and R3 is replaced with new resistors.
Q1 (2SK170BL) was after being removed from the PCB quickly swiped with a DMM (in diode mode) to chech for indication of faulty unit, with positive probe on middle pin and negative probe on Source and Drain reported back a 10ohm reading. Indicating a faulty transistor.
This test isn't conclusive but strongly suggest its shorted.
The circuit worked without issues (other than not tracking very well when adjusting P1/2 and offset=0,6ish V was more or less constant. I have previously biased this amp with no issues so I have some experience as to how it should react. The other channel biased nicely, however I stopped at 0,55V with 0,00V offset as further adjustment is fruitless without the other side working aswell.
I had reset P1/2 to zero ohm (measured over R5/6) twice and measuring probes were only relocated with power off. It was after I think the third reset R3 burnt instantly on power-up. Before the smell erupted each board played music, after smell only one side is good.
Is the 2SK170 that finicky?
My first attempt to bias this amp was in january (IIRC) and both circuits were fine, until something clicked knocking both channels out.
The circuits are untouched since then except for only replacing Q1-6 on both sides.
Will measuring each resistor (power off, in circuit) on the good side provide good data in order to evaluate the bad side?
Last question for now, since SK170 appear to be dead - the trannies are matched but what is their matching criteria? I mean, is the primary objective of matching them to have a matching J74/K170 pair for each side or is it worth while to have both pairs match each other? Mine were matched 8mA J74/K170 quad.
Last edited:
Finally back to my F5 and after some checking I was able to spot the problem. I had installed the body of the thermistor in contact with the central leg of the mosfet. Probably for wrong handling the insulation of the thermistor was damaged so I was shorting the negative input voltage directly to the output. Output was back to 0V as soon as I removed the thermistor from contact with the MOSFET leg.
Now a question before reinstalling the boards. My JFETs had idss in the order of 7.5 - 8 mA. Will the 2.2k value for R3 and R4 be enough or, since I've not yet reinstalled the board on the sink, shall I take the opportunity and replace them with a 3.3k or 4.75k? Thanks to all
Gior,
once you have this running you can adjust the VR (pot) to gradually bring up the voltage on the output FETs.
Then switch off and measure the resistance that you have across the VR and the other resistors.
Decide if the 3k3 is correct or could be lower value.
BTW (By The Way) you want a low current through the VR and most of the current through the fixed resistor/s. This condition requires that the fixed resistors be as low as possible and still turn on the output FETs to your chosen bias current.
once you have this running you can adjust the VR (pot) to gradually bring up the voltage on the output FETs.
Then switch off and measure the resistance that you have across the VR and the other resistors.
Decide if the 3k3 is correct or could be lower value.
BTW (By The Way) you want a low current through the VR and most of the current through the fixed resistor/s. This condition requires that the fixed resistors be as low as possible and still turn on the output FETs to your chosen bias current.
small effect due to excess noise in the trimmer.
small effect due to high tempco of the trimmer.
small effect due to variability of the wiper contact.
small effects due to trimmers in general.
Try to swamp the trimmer effects by minimising the proportion of current flowing through the trimmer.
small effect due to high tempco of the trimmer.
small effect due to variability of the wiper contact.
small effects due to trimmers in general.
Try to swamp the trimmer effects by minimising the proportion of current flowing through the trimmer.
For those who have set up a stock F5, I am curious of how much bias and offset varies.
I am not conserned, more curious.
Mine fluctuate +/-3 mV both on bias and offset. Typical or room for improvement?
DMM's are not that super precise either... (I have one channel up and running for the past 4 days and noticed this minute fluctuation (change is gradual over many seconds before it reverse)
I am not conserned, more curious.
Mine fluctuate +/-3 mV both on bias and offset. Typical or room for improvement?
DMM's are not that super precise either... (I have one channel up and running for the past 4 days and noticed this minute fluctuation (change is gradual over many seconds before it reverse)
Bias stability
Didn't see it in comments but I assume you are doing bias monitoring with the case top on (laid on, not screwed down). I monitor one bias point and output offset on each channel (need 4 meters) and note temperature at set points on heat sink with a cheap eBay remote temp monitor. 43C seems to be a good max spot for heat at outside points on the heatsink and puts the bias close to recommended value. Offset is very small and will drift a bit over longer periods. YMMV
Didn't see it in comments but I assume you are doing bias monitoring with the case top on (laid on, not screwed down). I monitor one bias point and output offset on each channel (need 4 meters) and note temperature at set points on heat sink with a cheap eBay remote temp monitor. 43C seems to be a good max spot for heat at outside points on the heatsink and puts the bias close to recommended value. Offset is very small and will drift a bit over longer periods. YMMV
I have 3DMM's on hand but only one channel running (waiting for replacement 2SK170) so I use 2 for N/P bias and one for offset.
Lid is on (not screwed) and with ambient temp=21degC I get 39degC on top of the fins (above fets 240/9240). Chassis is 4U/400mm.
(I expect a slight overall temp. increase when I get channel 2 up and running.)
I will also replace the 0,47R with a larger 5W component as these do run very hot @ .6V I can touch the FET's for second or two without feeling pain.
Last edited:
Why does it behave like Q1 is fully on: (gate drain shorted?)
- Are you realy shore the outputs are ok ?
- Are the protection transistors mounted the correct way and working ?
- Are / Have you checked resistor values for ( correct placement ) R19, R20 (and all the others).
- P1 P2 set to near zero at start up?
- Shorts due to soldering?
- example : without base resistor R14 output Q4 can turn on and therby frying R3 and possibly R4 and destroying Q1
Outputs checked and not swapped ?
Last edited:
Thank you for chiming in
Just a quick factual reply to your input:
1 - Yes. They are from a reputable supplier, tests ok and each amp (side) did play music before R3 gave up the smoke.
2 - Yes, see #1
3 - Yes, prior to populating the PCB each resistor was individually measured. The resistors of the good amp match the resistors of the ill amp.
4 - Yes, 0ohm measured over R5/R6 before adjusting.
5 - None found but will recheck every solder pad against Chassis GND.
I have also measured continuity between Audio GND and Chassis GND which read 16ohm (one CL60 used to lift Aud.GND, all Aud.GND share a single connection point.)
Speaker (+) terminals show no continuity and speaker (-) terminals show 16ohm. Same for signal input.
Since I did experience some anomalies during initial biasing of the ill amp it is possible Q4 was opened up inadvertably causing Q1 to fail, triggering the smoke of R3. The good amp have been cooking nicely for a week.
In a couple of days I'll have a couple genuine 2SK170BL's on hand and repopulate Q1 and start fresh with bulb tester and all
Tonite I soldered in the new 2SK170BL, wired it all up (single channel only) and tested with bulb tester. Check, dimmed all the way out.
Removed bulb tester, checked voltage which showed 52.4V (no load)
P1/P2 reset to zero ohm measured over R5/R6, and if maxed they stopped at approx 1k2ohm.
With power on offset was almost zero and started to adjust bias. After a few turns meters detected increased voltage and offset was countered etc, numbers added up. After a few more turns I stopped at .55V and offset=0.002V.
Lookimg good.
Powered off, connected the other channel and again inserted bulb tester when powering up both channels. The bulb tester did dim but not all the way out. Removed bulb tester.
Power up again with 1dmm on channel 1, the other over channel 2 and a 3rd at speaker out. Both sides started at 0,575v over the .47R and offset was good. Let it cook for an hour and when I returned channel 1 was cold and #2 was warm as expected.
Offset on #1 was now around 30mV but jumping about between 25-50mV. Bias measured zero. Further tweaking of P1/P2 gave no change in bias indication (yes, dmm was checked for proper contact) and offset was jumping, even with no change made to the pots.
Switched off, reset pots again but offset is still jumpy and I can not detect voltage over the source resistor (except a 50mV reading) offset unstable.
Measuring over R5/R6 showed a drift in resistance, slowly dropping.
No smoke detected, the bad channel biased up initially but failed some time during warm-up. I am cuurently clueless and hope for some guidance...
TIA, Haldor
Removed bulb tester, checked voltage which showed 52.4V (no load)
P1/P2 reset to zero ohm measured over R5/R6, and if maxed they stopped at approx 1k2ohm.
With power on offset was almost zero and started to adjust bias. After a few turns meters detected increased voltage and offset was countered etc, numbers added up. After a few more turns I stopped at .55V and offset=0.002V.
Lookimg good.
Powered off, connected the other channel and again inserted bulb tester when powering up both channels. The bulb tester did dim but not all the way out. Removed bulb tester.
Power up again with 1dmm on channel 1, the other over channel 2 and a 3rd at speaker out. Both sides started at 0,575v over the .47R and offset was good. Let it cook for an hour and when I returned channel 1 was cold and #2 was warm as expected.
Offset on #1 was now around 30mV but jumping about between 25-50mV. Bias measured zero. Further tweaking of P1/P2 gave no change in bias indication (yes, dmm was checked for proper contact) and offset was jumping, even with no change made to the pots.
Switched off, reset pots again but offset is still jumpy and I can not detect voltage over the source resistor (except a 50mV reading) offset unstable.
Measuring over R5/R6 showed a drift in resistance, slowly dropping.
No smoke detected, the bad channel biased up initially but failed some time during warm-up. I am cuurently clueless and hope for some guidance...
TIA, Haldor
I dont know if this is the problem with your amp, but I had a similar experience with F-5. Listening along, one channel goes dead. No smoke, no noise, just went gently quiet on one side. I measured nothing for bias. Measured nothing where power supply went into boards. Measured nothing where power came from rectifiers, set to AC nothing on other side either, all the way back to power cord. (Dual mono amp all the way back including 2 power cords) It turned out to be a bad connection where the wire connected to the Neutrik jack that cord connected to...do you have power to your cold board?
Who is reputable supplier?
Russellc
Last edited: