I hate shotgun replacement of parts, but sometimes there's no alternative.
I has a an amp problem that had huge crashing "static". Probing with 'scope made me suspicious of a transistor but nothing was obvious. Guess-work replacement of the transistor fixed it. Since you noted q7/q8 responded to cold-shot, they would come first in my swap tries. (You've probably discovered you can dispense cold-shot down the plastic accessory tube almost drop-by-drop for very localized testing.)
I know from frustrating experience how vexing these defects can be. Good luck!
I has a an amp problem that had huge crashing "static". Probing with 'scope made me suspicious of a transistor but nothing was obvious. Guess-work replacement of the transistor fixed it. Since you noted q7/q8 responded to cold-shot, they would come first in my swap tries. (You've probably discovered you can dispense cold-shot down the plastic accessory tube almost drop-by-drop for very localized testing.)
I know from frustrating experience how vexing these defects can be. Good luck!
Thank you! Looks like the only problem now is parts. I was cheap when ordering parts, and only ordered 1-2 extra of each transistors. In the case of the LTP q1/q2, I bought no extras as they were pricier. The LTP(q1/q2) also starts popping and static when heated, and then stops when cooled down. I'll tinkering around with them until I can get replacement parts.
Agreed, so far I've only changed D3. Same issue persists. q1/q2 I don't have spares. Will keep everybody posted.
Small problem... I only have 2 ksc1845 and 4 ksa992's. Looks like they're been marked as obsolete and at least the 1845's aren't in stock anywhere.
In practice the track between the earth terminals of the supply capacitors should be a thick and straight length of copper with nothing connected in between other than the transformer centre tap and the wire connecting to the chassis. Capacitor charging currents are substantial and rectification generates switching spikes. You don't want these flowing through low noise connections to earth.
All else should taken from an adjoining arm of of copper to form a T junction for other earth currents - starting with the most contaminated ones spaced a short distance down the T junction and the others spaced in order down to the quietest one - in this case G2. The circuit diagram shows a 5 Ohm resistor R4 in series with one of the two G2 earths. If this is really the 10 Ohm thermistor - used instead for safety reasons - it should be the last component at the far end the T junction.
In that case the RCA jack should be isolated from the chassis. I can see that to be the case in photos however the absence of isolating washers suggests otherwise. This raises the question is there an earth path between G2 and G1 on the pcb and there onward to the chassis via that route.
Thanks 🙂. This might take a while to get to Canada. I will try some swapping around with the ones I have to see if I can get any progress.
The rca jack connected is not the one originally used. The one that's in the chassis is heavier and is isolated from the chassis. I just have this one on there for now, because the noise isn't present without anything connected to the input terminal on the amp board. The popping and the static really point towards an issue with a semiconductor. It just seems that it is very sensitive and needs the input there for the conditions to trigger this situation.
That does not answer my question whether or not the input G2 earth in the circuit diagram is connected to earth. Often this would be via a separate wire unless G2 and G1 earth points are in series on the pcb. If the RCA socket you have been using is isolated from the chassis and there no link between G2 and G1 earth points on the board the input is floating.
Supply wiring and output wiring should be tightly twisted together so the ac electric fields around these more or less self cancel. It you separate them these fields will radiate into the spaces between which can induce noise into other conductors such other wiring including pcb tracks and transistor leads.
Another question to think about is the earth for your dc protection module - you mention this is floating why does it not have it's own dc earth point. This is parked right next door to your problem channel?
Last edited:
i see to it that at the rca jacks there is a terminating resistor right there, otherewise, wires can serve as antennae's....
No, the protection module is beside the good channel. Not sure where g1 and g2 connect. The guide describes the grounding as "A fully starred grounding scheme with the voltage stage and input stage "lifted" and brought back to the star ground" here is a picture that will help with pcb traces. 0v on the bottom right is input ground.
I'm satisfied that the input 0V connects to the Test Point 0V through traces on the pcb. All the same the In and 0V at the right hand side should be isolated from the chassis as you have said.
You could improve this by using a shielded cable for these connections and connect the screen wiring to the chassis to mitigate against the whole acting as antennae.
If you have some bad solder joints on this board a shotgun approach to replacing semiconductor components may reveal nothing. It would be simpler and a lot less work to measure the base and emitter voltages for each under power in separate operations.
Put the black lead of the meter into the speaker earth terminal and use the red lead to measure each of these two component leads - something you can do this with less risk of accidents and with more precision. The differences should be about 0.6V plus or minus depending on polarity.
If some discrepancy arises look at the components soldered to the respective parts to see if the solder work looks suspect and remedy that first.
No, the protection module is beside the good channel. Not sure where g1 and g2 connect. The guide describes the grounding as "A fully starred grounding scheme with the voltage stage and input stage "lifted" and brought back to the star ground" here is a picture that will help with pcb traces. 0v on the bottom right is input ground.
I have replaced q7/q8 and resoldered the transistors by the signal input and current mirror section. The noise is still there, but has changed; in that it takes longer to start, and lasts longer; but is less pronounced. I'm in a holding pattern of sorts, until I get more parts. I tested voltages on the transistors and everything looks fine.
Is there a test circuit I can build to scope the transistors under high voltage with a power supply and a signal generator? My power supply can go up to 60 vdc. Maybe if I test under high voltage and different frequencies I can get the issue to show up on the scope?
Is there a test circuit I can build to scope the transistors under high voltage with a power supply and a signal generator? My power supply can go up to 60 vdc. Maybe if I test under high voltage and different frequencies I can get the issue to show up on the scope?
Beautiful! Thanks. If I wanted to build this with an npn ksc1845, using 60 volts. What values would I need for the resistors and cap?
The ksc1845 is rated at 500mW, 120VDC and 50mA maximum, beta > 150.
With a 60VDC power supply, say we bias it at 28Vce, Ic = 5mA, and Pd =140mW.
Output DCV at collector wrt ground = 30V. The DCV at emitter = 2V.
Use 75V coupling capacitors of at least 1uF, and just test the operation at 1kHz.
ACV gain should be around x10.
Then we can use, with a back of envelope calculation:
Rc = 6.2k 1/2w
Re = 390R
R1 = 120k
R2 = 5.1k
Hazardous DC voltages are present in this circuit.
With a 60VDC power supply, say we bias it at 28Vce, Ic = 5mA, and Pd =140mW.
Output DCV at collector wrt ground = 30V. The DCV at emitter = 2V.
Use 75V coupling capacitors of at least 1uF, and just test the operation at 1kHz.
ACV gain should be around x10.
Then we can use, with a back of envelope calculation:
Rc = 6.2k 1/2w
Re = 390R
R1 = 120k
R2 = 5.1k
Hazardous DC voltages are present in this circuit.
Last edited:
Back in post 27 it was reported that when the Laboratory Power Supply was used to power the troublesome module you could not get a peep out of it (and apparently there were no signs of instability due to the reduced supply rails of that supply).
Collector to base capacitance in a transistor has an inverse relationship with the square root of voltage applied to the collector. There will be more collector to base capacitance with 30 volt supplies than with 60 volt ones.
Some adjustment to the stability compensation value might need to change but the thing held together well enough during this temporary test.
No doubt your equipment is housed in a metal case and at some distance from your amplifier chassis making it likely there is no radiated field intrusion into the set up of your amplifier in this situation. Don't think you are not exposed to this risk with your present amplifier power supply arrangements.
Your lab supply will be a professional piece of work. I suggest you take the covers off and look inside and take a lead from that.
Collector to base capacitance in a transistor has an inverse relationship with the square root of voltage applied to the collector. There will be more collector to base capacitance with 30 volt supplies than with 60 volt ones.
Some adjustment to the stability compensation value might need to change but the thing held together well enough during this temporary test.
No doubt your equipment is housed in a metal case and at some distance from your amplifier chassis making it likely there is no radiated field intrusion into the set up of your amplifier in this situation. Don't think you are not exposed to this risk with your present amplifier power supply arrangements.
Your lab supply will be a professional piece of work. I suggest you take the covers off and look inside and take a lead from that.
Last edited:
Thanks again, just tested 2 of them. The waveforms looked good. 3.8v in 56 Vpp out.
When I was building this amp, one of the channels had a meltdown and an output transistor shorted. Not sure if this was the channel, this is why I'm thinking it's a semiconductor issue instead of grounding. It's possible that grounding is an issue, but at this point I really think it's less likely. Also most of the top pads on this board don't connect to anything, they don't rely on the component lead as a via, the via's are soldered from the bottom and then run back to the bottom of other components. This amp was designed to be beginner friendly, which doesn't include being a master solder technician. I rarely have things fail due to my solder work.