Apparently the 10R emitter resistors burned.
I have used MJEs in all my V2 builds and VAS overcurrent has never been a problem.
What input transistors are you using?
Most probably there has been some inter-trace short.
If the 27K resistors are there then the VAS base current simply can't go high enough to increase VAS bias sufficiently and burn the 10R (needing more than 150mA for 10R/0.25w).
Check the 1N4148 diode, the input transistors and the VAS transistors if these are all ok. When powering up next time, use 1R resistors in both rails instead of one 10R in one rail. Trim the VAS bias and let the amp be "on" for a few minutes while carefully monitoring the VAS bias for any abnormal gradual increase with time.
If the 27K resistors are there then the VAS base current simply can't go high enough to increase VAS bias sufficiently and burn the 10R (needing more than 150mA for 10R/0.25w).
Check the 1N4148 diode, the input transistors and the VAS transistors if these are all ok. When powering up next time, use 1R resistors in both rails instead of one 10R in one rail. Trim the VAS bias and let the amp be "on" for a few minutes while carefully monitoring the VAS bias for any abnormal gradual increase with time.
That could be it.
IME, this happens when rail-resistors for bias measurement have high-ish and/or vastly different values (which it is in Prasi's case). This is why I recommended a 1R in each rail to test total bias.
This failure doesn't happened during test period but after a successful setup.
I can't see how 1R could help in this case.😕
Could be if it was triggered, usually by high frequency input signal. In my V2 experiments I have never seen this happen however and I always used 1R for measuring total bias and could not make the amp to oscillate or the VAS to cross-conduct even with 160KHz squarewave, with or without load.
This is why I specifically recommend that one does not insert measuring resistors into the supply rails.
Measure and adjust the output and driver stage bias currents with the full supply rail voltage applied.
With 1R in the rails the drop for 50mA draw is 50mV. I don't know how this could cause problems because I have never seen it do so as long as the resistors are less than a couple ohm in value and are same in both rails.
Could you please suggest a simpler/safer way to measure total amplifier DC current without using resistor voltage drop in the rails?
That is what is done to measure the VAS bias.
How to measure lateral MOSFET bias when they have no source resistors?
How to measure lateral MOSFET bias when they have no source resistors?
fit source resistors.
Helps to even out discrepancies between devices due to degeneration (NFB)
And helps with improving stability margins, both electrical and thermal.
Helps to even out discrepancies between devices due to degeneration (NFB)
And helps with improving stability margins, both electrical and thermal.
Aren't the HEXFETs more in need for source resistors?
Electrically and thermally the V2(that Prasi built) is as stable as the other single MOSFET O/P stage PeeCeeBees and the MOSFETs have no problem with uneven loading/dissipation with high load current, as my experiments showed.
Anyways, as there is no provision for source resistors in the PCB, I leave it to Prasi to decide whether to install them or not. Maybe this will solve the issue of VAS resistor burning (I would be the happiest if it does). 🙂
Electrically and thermally the V2(that Prasi built) is as stable as the other single MOSFET O/P stage PeeCeeBees and the MOSFETs have no problem with uneven loading/dissipation with high load current, as my experiments showed.
Anyways, as there is no provision for source resistors in the PCB, I leave it to Prasi to decide whether to install them or not. Maybe this will solve the issue of VAS resistor burning (I would be the happiest if it does). 🙂
I can't answer that since I am not an amplifier Designer.
Lateral mosFETs have that negative tempco once the current is above ~100mA and that helps a lot with thermal stability.
One of our Members tried to blow up an output stage by deliberately drawing far in excess of rated power and the laterals seemed to enter some fail safe operating mode and refused to die.
Lateral mosFETs have that negative tempco once the current is above ~100mA and that helps a lot with thermal stability.
One of our Members tried to blow up an output stage by deliberately drawing far in excess of rated power and the laterals seemed to enter some fail safe operating mode and refused to die.
I don't know if you are an amplifier Designer or not but obviously you have a long experience with amplifiers and your suggestions have helped a lot of DIYers in this forum including myself.
I just reported my DIY findings, that includes parallel HEXFETs without source resistors blowing violently but the Laterals do indeed seem to have some secrets, reason behind me not letting them go off PeeCeeBee. 🙂
I just reported my DIY findings, that includes parallel HEXFETs without source resistors blowing violently but the Laterals do indeed seem to have some secrets, reason behind me not letting them go off PeeCeeBee. 🙂
I repeat again,Prasi's amplifier passed the setup test successfully,play well and then burned.
It seems like we don't understand each other....🙁
Prasi are you sure about your supplier?
How possible is your transistors are fake?
It seems like we don't understand each other....🙁
Prasi are you sure about your supplier?
How possible is your transistors are fake?
Last edited:
I have a commercial power amplifier (German manufacture, not China) and I touched the unconnected input with an RCA interconnect and it instantly blew up.
Sent for repair under warranty.
After it came back, I stupidly touched that unconnected input with an oscilloscope probe and it instantly blew up.
Some designs are very intolerant of some operating conditions.
One should be testing to find and eliminate these intolerant behaviours.
Sent for repair under warranty.
After it came back, I stupidly touched that unconnected input with an oscilloscope probe and it instantly blew up.
Some designs are very intolerant of some operating conditions.
One should be testing to find and eliminate these intolerant behaviours.
Last edited:
Hi Thimios.
Please see Post#2507.
Let's wait for Prasi's tests with two 1R rail resistors. I hope his MJEs are genuine (I have heard of fakes of them getting into market).
I have done the same test repeatedly on all PeeCeeBees and the output goes completely insane but nothing blew up (except the zobel resistor on one occassion). So I guess PeeCeeBee is fairly tolerant of this kind of abuse.
Hello Thimios and Shaan,
I had been using the 2SK/2SJ in DIY Apex FX-8 stereo for a long time without any problems. Since I didnt have extra pair, I cabbibalized the FX-8 and used the laterals in PeeCeeBee .
I have measured the hfe of BC's (matched from about 10 nos each). Also measured the Hfe of MJE's, though didnt try to match. I get that transistors showing Hfe reading are atleast in working condition and not dead.
Supplier , although local, never had a problem with transistors.
Another thing is that I used 10 R rail resistor on both rails, both rails had approx the same reading so posted only one.
offset initially was something like 110mV (the pots were not set to same value ). SO had to adjust the pots to bring down the offset to about 1.2 mV. (I adjusted the pots on both sides approximately equally.
the front end +VAS current, current thro VAS emitter resistor (and hence the
current thro i/p pair) was approximately same within +/- 0.1-0.2 mA, so didnt record the reading on negative rail.
I hope this helps in finding the problem.
Also after soldering, I carefully checked the bottom (with a led torch from the component side), there were no shorts.
after testing, I had connected to LFE of a home theater receiver with a cross over of 100Hz. I worked for about 1 hour before blowing up. earlier under test, it had worked for about 1 hour in idle condition and playing music thro a test speaker.
Last edited: