Thank you Fab and ggetzoff – I have something to work with now, and a little more hope to recover.
Just to confirm:
- R15 and R16 are 0.22 ohm in my case. I could not get 0.2, so I got 0.22 ohm 3W resistors and they actually measured 0.23 ohm on my cheap component tester. My DMM is not accurate enough for such low resistance. I used the component tester also to measure hFE and match Q’s 1 to 4 (BC547 and BC557)
- I will check again, but almost 100% sure that Q7 and Q8 are in the correct positions
- R4 is mounted, but hidden behind the C1 caps. Yes, there are 2 x 110pF caps in parallel. I could not get 220pF mica caps at the time
- Measurements went well with sections 8.7 Current source adjustment and 8.8.5 Adjust current in input stage.
- Up to this point only V+, V- and PGND was supposed to be connected. I did not connect PGND, so only V+ and V-
- My problem (smoking resistors R15 and R16) started when I connected V+, V- and GND to attempt measurements for section 8.17.1 Adjust output Mosfet current close to intended final bias. At this point I should have skipped to Section 20.1.3.3 of Addendum 2.
- Maybe my problem is that I don’t understand the difference between GND and PGND in this case? Maybe I switched them around? I assumed that for a dual polarity PSU one refers to the outputs as V+ / GND / V-, and that PGND is only connected to the earth on the supply and to the wall socket in case of a short circuit?
I will unfortunately not be able to attend to this for the next two or three weeks.
Just to confirm:
- R15 and R16 are 0.22 ohm in my case. I could not get 0.2, so I got 0.22 ohm 3W resistors and they actually measured 0.23 ohm on my cheap component tester. My DMM is not accurate enough for such low resistance. I used the component tester also to measure hFE and match Q’s 1 to 4 (BC547 and BC557)
- I will check again, but almost 100% sure that Q7 and Q8 are in the correct positions
- R4 is mounted, but hidden behind the C1 caps. Yes, there are 2 x 110pF caps in parallel. I could not get 220pF mica caps at the time
- Measurements went well with sections 8.7 Current source adjustment and 8.8.5 Adjust current in input stage.
- Up to this point only V+, V- and PGND was supposed to be connected. I did not connect PGND, so only V+ and V-
- My problem (smoking resistors R15 and R16) started when I connected V+, V- and GND to attempt measurements for section 8.17.1 Adjust output Mosfet current close to intended final bias. At this point I should have skipped to Section 20.1.3.3 of Addendum 2.
- Maybe my problem is that I don’t understand the difference between GND and PGND in this case? Maybe I switched them around? I assumed that for a dual polarity PSU one refers to the outputs as V+ / GND / V-, and that PGND is only connected to the earth on the supply and to the wall socket in case of a short circuit?
I will unfortunately not be able to attend to this for the next two or three weeks.
I was in the same boat. The build guide has all the details, but they were not clear enough for me. I guess you just connect both GND and PGND to ground, and that's it.
Fab may have more details, if necessary.
I think per Fab and the manual, one of the GND or PGND has a floating ground via a resistor. Which means that you should connect both the grounds to the PSU ground for proper measurement and setup. One of my channels blew up after an year of play because the PGND spade cable because of frequent plugging and unplugging had come out loose and the adjacent resistor blew up. Based on Fab's suggestion now I have made sure that I get a slightly higher wattage 2w resistor and also made sure that the ground cables from the PSU board are tight fit.
Best of luck.
Best of luck.
Gnd and PGND
See below extract of manual where I just manually added the isolation R between input/feedback and GND pin. Then you need to connect GND and PGND at the power supply central ground connection. During testing, you can omit PGND in the first steps of the manual but not when you go back. Therefore, I recommend that you always connect both GND and PGND pins to the PSU gnd.
Fab
See below extract of manual where I just manually added the isolation R between input/feedback and GND pin. Then you need to connect GND and PGND at the power supply central ground connection. During testing, you can omit PGND in the first steps of the manual but not when you go back. Therefore, I recommend that you always connect both GND and PGND pins to the PSU gnd.
Fab
Attachments
Hi Harry3
These are good but they take twice the space on the pcb….🙄
It would probably need an increase of USSA5 pcb size….
Fab
It looks like I have good news and bad news. The good news – my second 5B board seems fine and should bias correctly. First want to upgrade the R15 and R16 resistors to 5W. The bad news - on the first board it looks like I blew up the two output transistors (2SA1943, 2SC5200). Waiting for replacements. Time will tell if they are the only damaged components. Will try to replace them without lifting everything off the heatsinks.
I also decided to kick this habit of blowing up amps from now on.
I also decided to kick this habit of blowing up amps from now on.
Hi Twocents
Glad about the good news.🙂
For the bad one, the build manual has been written in steps so you can test and validate after each test step (you will not see that very often, if ever). So you simply remove the failed output bjt and retest the previous steps so when you installed the replaced bjt properly then it will work for sure.
As a precaution, either always use a lightbulb or variac on AC power or small current fuses on each DC rail. I use 2A fuses on each rail during my bench test and never damaged any power transistors even when doing the wrong thing….
Good luck and I support you in your resolution😛
Fab
Glad about the good news.🙂
For the bad one, the build manual has been written in steps so you can test and validate after each test step (you will not see that very often, if ever). So you simply remove the failed output bjt and retest the previous steps so when you installed the replaced bjt properly then it will work for sure.
As a precaution, either always use a lightbulb or variac on AC power or small current fuses on each DC rail. I use 2A fuses on each rail during my bench test and never damaged any power transistors even when doing the wrong thing….
Good luck and I support you in your resolution😛
Fab
Hi AnthonyA
Not yet
I need to finish 5.2 version and then 3.2b. It will be shortly when I have the chance.
Fab
Not yet
I need to finish 5.2 version and then 3.2b. It will be shortly when I have the chance.
Fab
Hi Fab,
I finally continued working on the USSA-3, I'm now testing the input stage.
The manual states:
8.1.2 Adjust voltage in input stage
Connect V+ and V- and PGND (the one close to V-) to power supply (from 15Vdc to 30Vdc).
Adjust P2 for about TBD between TP8 and V-.
Adjust P1 for about TBD between TP7 and V+.
What's the voltage drop (TBD) I should aim for
over the resistors between TP7, TP8 and power supply voltage
or over the input transistors ?
I have a 24vdc PS.
I have the manual "USSA-33 AMPLIFIER BOARD Fab rev06p"
I just see that the manual also states "TBD" in the next testing phases.
Can you give me the voltages/currents the transistors of each stage(input,driver,output) should see ?
I finally continued working on the USSA-3, I'm now testing the input stage.
The manual states:
8.1.2 Adjust voltage in input stage
Connect V+ and V- and PGND (the one close to V-) to power supply (from 15Vdc to 30Vdc).
Adjust P2 for about TBD between TP8 and V-.
Adjust P1 for about TBD between TP7 and V+.
What's the voltage drop (TBD) I should aim for
over the resistors between TP7, TP8 and power supply voltage
or over the input transistors ?
I have a 24vdc PS.
I have the manual "USSA-33 AMPLIFIER BOARD Fab rev06p"
I just see that the manual also states "TBD" in the next testing phases.
Can you give me the voltages/currents the transistors of each stage(input,driver,output) should see ?
Hi Danny
You have the partial manual. I sent you earlier today the full version. You should find your answers there and if not do not hesitate to chime in here.
good luck
fab
Thanks Fab!
Manual received and continued with setting the driver stage.
For my build I would move the output transistors a little further from the PCB, this for better heat distribution.
Will this give a problem, since the gate resistor will be at 5cm ?
Or should I move the gate resistor and the zener closer to the output transistor?
Manual received and continued with setting the driver stage.
For my build I would move the output transistors a little further from the PCB, this for better heat distribution.
Will this give a problem, since the gate resistor will be at 5cm ?
Or should I move the gate resistor and the zener closer to the output transistor?
Just move the gate resistor and Zener to the pins of the transistors and you're good. I did the same, see here: https://www.diyaudio.com/community/threads/ussa-5-build-with-review.323778/post-6451152
Hi Danny
yes, Mbrennwa implementation is quite a nice integrated one👍
In its case for the USSA-5 - with longer wires than 5 cm - I remember that there was no sign of oscillation (with gate resistors and zener on mosfet pins). However, the USSA-5 has a zobel at the output where the USSA-3 had none since not found necessary. But since you plan to have shorter wires and with resistors (and zener) installed on mosfet gate ( and close to transistor body) there should be no issue.
Normally the position of output transistors on a 300mm length heatsink is adequate for proper heat distribution so I suppose you have a different heatsink length or you want to get perfect optimization…
I note that you have long leads for the thermistors which is not necessary and is not best practice. They only need to take heatsink temperature and not instantaneous ones of the mosfet. Half actual lead length was used successfull.
I also note that the pot P1/P2 are installed reversed (screw) so turning right will decrease the output bias current instead of increasing. Just be aware of it while doing the adjustment.
The drivers body is isolated so no need for insulation between them and heatsink.
For input stage can you tell your jfet IDSS current you have used.
fab
yes, Mbrennwa implementation is quite a nice integrated one👍
In its case for the USSA-5 - with longer wires than 5 cm - I remember that there was no sign of oscillation (with gate resistors and zener on mosfet pins). However, the USSA-5 has a zobel at the output where the USSA-3 had none since not found necessary. But since you plan to have shorter wires and with resistors (and zener) installed on mosfet gate ( and close to transistor body) there should be no issue.
Normally the position of output transistors on a 300mm length heatsink is adequate for proper heat distribution so I suppose you have a different heatsink length or you want to get perfect optimization…
I note that you have long leads for the thermistors which is not necessary and is not best practice. They only need to take heatsink temperature and not instantaneous ones of the mosfet. Half actual lead length was used successfull.
I also note that the pot P1/P2 are installed reversed (screw) so turning right will decrease the output bias current instead of increasing. Just be aware of it while doing the adjustment.
The drivers body is isolated so no need for insulation between them and heatsink.
For input stage can you tell your jfet IDSS current you have used.
fab
Last edited:
Yes, it's a 400mm length heatsink, so the transistors are a little further.
OK, some work for my desoldering pump 🙂
I'll put the gate resistors and the zeners close to the transistors and I will shorten the leads of the thermistors.
I even didn't notice that the pots are reversed installed, they can stay like this 🙂
For a 2.5v drop over the 2sk170 resistor/thermistor/pot I measured 650ohm, that's 3.85ma
For the 2sj74 I measured 680ohm, that's 3.68ma
OK, some work for my desoldering pump 🙂
I'll put the gate resistors and the zeners close to the transistors and I will shorten the leads of the thermistors.
I even didn't notice that the pots are reversed installed, they can stay like this 🙂
For a 2.5v drop over the 2sk170 resistor/thermistor/pot I measured 650ohm, that's 3.85ma
For the 2sj74 I measured 680ohm, that's 3.68ma
Hi Danny
you have measured ID and IDSS is measured with source resistor as 0 ohms. Your extrapolated IDSS is on the lowish side so it is fine to get slightly higher damping. Keep on the good work so far!
Fab
you have measured ID and IDSS is measured with source resistor as 0 ohms. Your extrapolated IDSS is on the lowish side so it is fine to get slightly higher damping. Keep on the good work so far!
Fab