Have a look here XRK,
Q5 should not be located on top of Q7, only a spacer is shown on top of the 2SC5200 (Q7)
See here post 561 from Prasi.
http://www.diyaudio.com/forums/soli...-quasi-complimentary-mosfet-amplifier-57.html
BR,
Eric
Q5 should not be located on top of Q7, only a spacer is shown on top of the 2SC5200 (Q7)
See here post 561 from Prasi.
http://www.diyaudio.com/forums/soli...-quasi-complimentary-mosfet-amplifier-57.html
BR,
Eric
I double checked the BD139. It is correctly wired on both boards as you can see and the difference in wiring color between two boards is what caused the confusion but I was not concentrating on making color of 3 wires consistent between amps but just the orientation is correct. I know having that flipped the bias won't adjust and the amp probably pegs one rail.
Here is
Here is
Attachments
It's not - I am just using a BD139/140 as a convenient perfectly sized spacer on top of the big BJT.
Hallo Friends.
Eric and Patrick next week i can send you a pair boards.
Send a mail please and it goes on way.
Bangla.
Hi Bangla H,
Just received my pair of boards, thank you very much !!
Don't expect my boards to be complete in a few days like XRK and Terry..lol..I'm a slow builder
Thanks,
Eric
inductor is simple to make if you have insulated (enamaled) copper wire (magnet wire, I think its called). 1.2mmdia wire , 0.75inch length of coil and 0.5inch dia of coil should give you approx 1.5uH.
Air Core Inductor Coil Inductance Calculator
I use an AA battery to wind mine, approx 12-15 turns should be sufficient.
Last edited:
Unfortunately, you can't buy stuff like this. There are a few items that are close (speaker crossover air core coils) but they will be too big to fit on the board. Get yourself some enameled magnet wire (I use 18ga) and wrap 12 to 15 turns around a AA battery. Sandpaper the ends so you can solder it. Some people make the coil coaxial with the resistor inside. I think that can change the L value and may induce currents in the resistor so I have them perpendicular to each other.
Edit: I see Prasi said basically the same thing LOL as we cross posted.
I had been running on 2 or 3 hrs sleep so thought maybe clearer thinking would result from a good break. Anyhow came back with fresh eyes and spent last hour retouching and reflowing all solder joints with my iron and using magnifier goggles to carefully examine all joints and pad for lifted traces/bridges etc. All looks good so far. About to go in and try to re-measure all voltages requested by Aksa, this time with a reasonable bias current of 80mA and using the 33k (32k actual value) at R5. Will report back soon with voltages.
I am listening to my FH9 as I work on these repairs, and it sure sounds good and fired up perfectly the first time. This amp, on the other hand, is a bit of high maintenance lady.
Hello Hugh:
I found out that my source used for testing the amps was bad and connected a different DAC today. It plays music nicely even with -2.5v offset. I listened for a while and the sound is clear and kind of a "sweet" tone. The bass is very good - felt like a good deep reach.
Then I went to measure the following on a live circuit... and of course, had an accident...
1. Voltage across Vbe multiplier, Q4, from C to E of BD139. Should be 1.90 volts
2. Voltage across LED and series 4004 diodes. Should be 3.25 volts
3. Voltage from gate of Q6 nmos to base of inverter, Q5. Should be 5.19 volts.
4. Voltage across R20, for 0.47R this should be around 50mV
5. Voltage at output wrt ground, should be 0mV but I think it's about +2? (positive?)
6. Voltage wrt ground at left side of VR1, should be -3.2V.
7. Voltage wrt ground at right side of VR1, should be -3.82V.
8. Voltage at wiper of VR1, and voltage at base of Q1, all wrt ground.
1. XRK: remeasured at 1.196v for 103mA bias current
As I rotated the BD139 back in place after measuring, the B pin touched a protrusion connected to +35v rail. The Blue LED died. I went back and replaced the blue LED and it lights up but the amp pegs the rail and draws 17v across the 10R safety resistor. Without the safety resistor, the IRFP got very hot quick - I could smell stuff coming off I assume either the IRFP240 is dead and shorted, or something is driving its gate wide open. Although the IRFP240 doesn't measure 0ohms across pins - still in the ten kohm to Mohm range when installed Something else died.
2. voltage across 2 red LEDs is 3.50v
XRK: remeasured at 3.49v across red LEDs
HD: This gives us a total bias voltage of 2.13+ 3.50= 5.63. This is a heavy bias; you could increase the resistance in the BASE to EMITTER of the Vbe multiplier so that the voltage across the entire shebang reduces to Thimios' value, that is 5.19V rather than 5.63V. This should reduce the idle current around 100mA, and if it's 400mA it should be then 300mA, but to be honest this is a bit too high. I would not use more than 150mA. Even then it will be pretty warm, this is 0.15 x 70 = 10.5W, about what it should have.
XRK: I reset to 23mV for 0.22R
3. nmos G (-) tp Q5 base (+) is -5.06v
XRK: could not get to this as amp died with accident above.
HD: This confuses me. What is the voltage from GATE of nmos Q6 to BASE Q5 pnp? It should be 5.20V, no more.
4. R20 voltage is 90mV across 0.22R (I adjusted this for 400mA current).
Normally I run 80mA but trying to see if XO distortion goes away - no.
HD: 90mV across 0.22R is 410mA and you will be dissipating around 28.7W across your heatsink for each channel. This is far too high and you need to reduce your bias generation by increasing Vbe resistance on the Vbe multiplier as above.
5. -1.3v is lowest I can get it
XRK: Now the lowest is -2.50v with the 32k at R5
HD: You need to put the black probe on the ground, negative probe is always on GROUND, and red probe on the output. If you do this should get +1.3, not negative, and the next question is this: CAN YOU GET THIS VOLTAGE UP CLOSE TO ZERO?
XRK: no it will not go any closer to zero.
6. -4.76v
XRK: could not repeat this as amp died
7. -3.50v (side facing top edge where Fastons are)
XRK: could not repeat this as amp died
HD: This is wrong way around. I should read -3.5V (black on grnd, red on left side of VR1 near leds) and -4.76V wrt ground on the RIGHT side of the VR1, near 10k 1W R6. Is this correct?
8. voltage at wioer of VR1 is -3.50v
XRK: could not repeat this as amp died
HD: Have your removed R4, and replaced R5 with 33k? What is the voltage, with black on gnd and red on base of Q1, wrt ground now?
[/QUOTE]
XRK: Yes, R4 gone (otherwise dc offset would not respond. R5 is now 32k. Could not get to Q1 measurement as amp died.
If you can tell me what I need to replace now to get the amp working again after basically the B pin of Q4 touched +35v rail.
So, now I can say the amp is capable of playing music nicely. The DC offset just doesn't have enough compliance with anything I do: 100k, 32k, 51k, or 100k trimpot at R5. What is need to drop the offset to zero?
I found out that my source used for testing the amps was bad and connected a different DAC today. It plays music nicely even with -2.5v offset. I listened for a while and the sound is clear and kind of a "sweet" tone. The bass is very good - felt like a good deep reach.
Then I went to measure the following on a live circuit... and of course, had an accident...
1. Voltage across Vbe multiplier, Q4, from C to E of BD139. Should be 1.90 volts
2. Voltage across LED and series 4004 diodes. Should be 3.25 volts
3. Voltage from gate of Q6 nmos to base of inverter, Q5. Should be 5.19 volts.
4. Voltage across R20, for 0.47R this should be around 50mV
5. Voltage at output wrt ground, should be 0mV but I think it's about +2? (positive?)
6. Voltage wrt ground at left side of VR1, should be -3.2V.
7. Voltage wrt ground at right side of VR1, should be -3.82V.
8. Voltage at wiper of VR1, and voltage at base of Q1, all wrt ground.
1. XRK: remeasured at 1.196v for 103mA bias current
As I rotated the BD139 back in place after measuring, the B pin touched a protrusion connected to +35v rail. The Blue LED died. I went back and replaced the blue LED and it lights up but the amp pegs the rail and draws 17v across the 10R safety resistor. Without the safety resistor, the IRFP got very hot quick - I could smell stuff coming off I assume either the IRFP240 is dead and shorted, or something is driving its gate wide open. Although the IRFP240 doesn't measure 0ohms across pins - still in the ten kohm to Mohm range when installed
Something else died.2. voltage across 2 red LEDs is 3.50v
XRK: remeasured at 3.49v across red LEDs
HD: This gives us a total bias voltage of 2.13+ 3.50= 5.63. This is a heavy bias; you could increase the resistance in the BASE to EMITTER of the Vbe multiplier so that the voltage across the entire shebang reduces to Thimios' value, that is 5.19V rather than 5.63V. This should reduce the idle current around 100mA, and if it's 400mA it should be then 300mA, but to be honest this is a bit too high. I would not use more than 150mA. Even then it will be pretty warm, this is 0.15 x 70 = 10.5W, about what it should have.
XRK: I reset to 23mV for 0.22R
3. nmos G (-) tp Q5 base (+) is -5.06v
XRK: could not get to this as amp died with accident above.
HD: This confuses me. What is the voltage from GATE of nmos Q6 to BASE Q5 pnp? It should be 5.20V, no more.
4. R20 voltage is 90mV across 0.22R (I adjusted this for 400mA current).
Normally I run 80mA but trying to see if XO distortion goes away - no.
HD: 90mV across 0.22R is 410mA and you will be dissipating around 28.7W across your heatsink for each channel. This is far too high and you need to reduce your bias generation by increasing Vbe resistance on the Vbe multiplier as above.
5. -1.3v is lowest I can get it
XRK: Now the lowest is -2.50v with the 32k at R5
HD: You need to put the black probe on the ground, negative probe is always on GROUND, and red probe on the output. If you do this should get +1.3, not negative, and the next question is this: CAN YOU GET THIS VOLTAGE UP CLOSE TO ZERO?
XRK: no it will not go any closer to zero.
6. -4.76v
XRK: could not repeat this as amp died
7. -3.50v (side facing top edge where Fastons are)
XRK: could not repeat this as amp died
HD: This is wrong way around. I should read -3.5V (black on grnd, red on left side of VR1 near leds) and -4.76V wrt ground on the RIGHT side of the VR1, near 10k 1W R6. Is this correct?
8. voltage at wioer of VR1 is -3.50v
XRK: could not repeat this as amp died
HD: Have your removed R4, and replaced R5 with 33k? What is the voltage, with black on gnd and red on base of Q1, wrt ground now?
[/QUOTE]
XRK: Yes, R4 gone (otherwise dc offset would not respond. R5 is now 32k. Could not get to Q1 measurement as amp died.
If you can tell me what I need to replace now to get the amp working again after basically the B pin of Q4 touched +35v rail.
So, now I can say the amp is capable of playing music nicely. The DC offset just doesn't have enough compliance with anything I do: 100k, 32k, 51k, or 100k trimpot at R5. What is need to drop the offset to zero?
Last edited:
XRK,
I will think about all this over my dinner, but if you can set the quiescent with 23mV over 0.22R on the source resistor on the IRFP240 we can say:
1. The mosfet is OK, no damage.
2. Very likely the C5200 quasi device is very likely OK too.
3. If the offset is within 5V of positive or negative wrt ground, then the issue is merely the bias system.
Now, some fixes:
Put heatshrink on ALL of the probe ends, and heat them on. Then using a Stanley, cut off the last 2mm of the probe so the metal is then exposed. Then you do not touch other parts when you home in on the tip of the probe. Watch for the side of car as you drive though a gate.......
Now, using the DMM, measure all base to emitter on ALL bipolar transistors. They should ALL measure somewhere between 0.55 and 0.7 volts. No exceptions.
Finally, measure the voltage from gate of the nmos to source of the nmos. It should be around 4.15 volts, positive on the gate, negative on the source.
Lastly, get rid of the 100k at R5 and replace with 33k. This screws with the bias voltage and louses up the offset voltage.
Conclusion: You had a working amplifier and will again. Replace the BD139 Vbe multiplier and the voltage amplifier whose 10R emitter is connected to the negative rail. This will restore function to the damaged section. Now you have only measurement, and detail to get the offset in check.
Cheers,
Hugh
I will think about all this over my dinner, but if you can set the quiescent with 23mV over 0.22R on the source resistor on the IRFP240 we can say:
1. The mosfet is OK, no damage.
2. Very likely the C5200 quasi device is very likely OK too.
3. If the offset is within 5V of positive or negative wrt ground, then the issue is merely the bias system.
Now, some fixes:
Put heatshrink on ALL of the probe ends, and heat them on. Then using a Stanley, cut off the last 2mm of the probe so the metal is then exposed. Then you do not touch other parts when you home in on the tip of the probe. Watch for the side of car as you drive though a gate.......
Now, using the DMM, measure all base to emitter on ALL bipolar transistors. They should ALL measure somewhere between 0.55 and 0.7 volts. No exceptions.
Finally, measure the voltage from gate of the nmos to source of the nmos. It should be around 4.15 volts, positive on the gate, negative on the source.
Lastly, get rid of the 100k at R5 and replace with 33k. This screws with the bias voltage and louses up the offset voltage.
Conclusion: You had a working amplifier and will again. Replace the BD139 Vbe multiplier and the voltage amplifier whose 10R emitter is connected to the negative rail. This will restore function to the damaged section. Now you have only measurement, and detail to get the offset in check.
Cheers,
Hugh
Last edited:
Aksa,
I just pulled all transistors off except 5200 and Q5, and measured them. All seem ok - Hfe's register and on IRFP240, G-D and G-S is "open" and S-D is 6Mohm. I was able to set bias current before the accident. The short was due to partially exposed leg of Vbe multiplier as I was moving it around under the board. But yes, I should have been more careful. It's late here and will take up again tomorrow night. Thanks for your help.
The detail is why the DC offset is so far off and what can be used to move it?
I just pulled all transistors off except 5200 and Q5, and measured them. All seem ok - Hfe's register and on IRFP240, G-D and G-S is "open" and S-D is 6Mohm. I was able to set bias current before the accident. The short was due to partially exposed leg of Vbe multiplier as I was moving it around under the board. But yes, I should have been more careful. It's late here and will take up again tomorrow night. Thanks for your help.
The detail is why the DC offset is so far off and what can be used to move it?
In studying the schematic diagram for this amp, it seems to me that the DC offset wil be affected by the voltage drop across the red LEDs. An offset of -2.50v after a gain of 31 (1500R/47R) is caused by 78mV offset at input to Q1. I feel like my LEDs are kind of an intense high brightness red and their voltage drop appears to be higher at 3.50v vs the recommended 3.2v. Could this be the cause? I have several other colors to work with to get a lower drop. Even a single green that is supposedly 3.2v. Perhaps concentrate on the amp front end to get rid of this DC offset issue. This is a singleton input amp and it's not possible to put a trim pot between the legs of a traditional two transistor LTP to get rid of DC offset. Am I on the right track?
Aksa, when you say voltage drop across LEDs and 1N4004 should be 3.25v that means LEDs alone should be 3.25-0.56v or 2.70v. I am measuring 3.5 is across the 2 red LEDs. Can you tell me exactly the test points to measure this voltage drop? Do you mean on the signal ground side of C2 and pin 3 of VR1?
Aksa, when you say voltage drop across LEDs and 1N4004 should be 3.25v that means LEDs alone should be 3.25-0.56v or 2.70v. I am measuring 3.5 is across the 2 red LEDs. Can you tell me exactly the test points to measure this voltage drop? Do you mean on the signal ground side of C2 and pin 3 of VR1?
If two red LEDs have too high a Vf then try replacing one or both with IR LED. 1.75Vf+1.3Vf=3.05V
Or replace one with a signal diode.
1.75Vf +0.75Vf = 2.5V
Between Schottky diodes, signal diodes, IR LEDs, red LEDs, yellow LEDs and green LEDs, you have an enormous range of compound possibilities from ~1V to ~4V in small increments.
Or replace one with a signal diode.
1.75Vf +0.75Vf = 2.5V
Between Schottky diodes, signal diodes, IR LEDs, red LEDs, yellow LEDs and green LEDs, you have an enormous range of compound possibilities from ~1V to ~4V in small increments.