1: Usually is not good to use component as mechanical distance, thermal expansion make the lead break at case in long time. 2: Another bad thing that usually should be avoid is a big hole in middle of pcb , that hole will be in ground plane , and different frequency may go different path . 3: I usually use two transistor as thermal sensor 4: clipping show that transistor get saturated and the don't come out , you can put baker diode in VAS so you get headroom. Another test to find out which one is saturated is put noise signal ( 20k limited ) and watch for top if they can come out from saturation. 5: phase margin and gain margin if you have networks analyser 6 : cut transistor lead (big one), wire it to pcb . 7: put 100p cap parallel wirh feedback resistor
I agree with this idea, especially since the output transistors are on opposite side of the PCB. Put two thermal sense transistors in your design, one on each side of the PCB near the output transistors.
I have been a sub for many years, and this is a great project to build onto your channel. I am keenly looking forward to seeing how this develops and hopefully, there will be an option to buy the kit sometime soon.
I got the same type of transistors! Toshiba ( 2SA1943 and 2SC5200 ) 10 of each and I really wanted to make a nice amplifier with these so I hope your project works great! Great video as always!!!👍
The output transistors are 12MHz vs. 30MHz in the JAT501, and the cob is also about half that of the transisitors used in the 501. I don't know if that's the issue, but changing outputs in an amplifier, sometimes has the weirdest effects.
I've been thinking along the same lines. He used old transistors not the usual sc5200 and sa1943. I'm not convinced this is a failure at all. keep pluggin John.
I wonder if it might be useful to do some open loop tests. I think it can be hard to say for sure that it is the VAS stage when the loop is closed. Remove the feedback resistor R12 and put a very small signal in to see of the VAS stage is clipping that way without feedback. Of course you probably want to run it with very low voltage rails to keep from blowing things in open loop mode. I bet it is the output transistors being too slow that is the cause.
Perhaps an easier way to measure the open loop performance would be to short out R13 so that you were just getting DC feedback. That would be easy to do as a temporary step as you would just have to add a jumper and you wouldn't have to deal with small dc offsets on your input test signal.
I purchased a ljm20.5 amp module off ebay and it oscillates like crazy, it unusable, I had to add miller caps and a output low pass, just to get it to work, it still just slightly unstable.
I'd be interested to see how it performed if you removed one of the output transistors off each rail. This is just academic interest. Obviously doing this would defeat the aim of creating an amplifier with higher output than the JAT501.
John, I noticed a (barely audible) high frequency when the camera is approaching the board, for example @14:46. I wonder if it is related to your oscillation problem. Sounds to me like an interference from a nearby device.
I don't like that asymmetric ringing that the 501 also has, I find these kind of issues very hard to trace because the whole amp is within a feedback loop I can't tell the difference between cause and effect, think I would blindly just go in and add 22pf ceramic caps across base, collector all transistors and if that cures it start removing them to see which transistor is the cause. I'd Like to know if there's a logical method though
I was wondering how the poles of the output might have different stability. While circuit board layout is a possibility, I think you may have hit the nail on the head. I understand it is actually physically impossible to create perfectly matched pnp and npn transistors. If memory serves, it wasn't until the mid 70s that closely matched pairs were manufactured.
@@Roger_Gadd those output transistors were from the same amp so unless he's mixed left and right pairs they are matched so I'm thinking as you say pcb layout or it's predrivers but I'm from the repair side of electronics not design so this is outside of my comfort range
@@darrenmurphy6251 I confess to being an amateur, but I enjoy electronics and learning from John's videos. From my personal study of transistor theory, I understand that although the gains of a pnp-npn pair may be matched, one or more of the internal capacitances, resistances, and time delay of the opposing transistors won't match. The output transistors used in this prototype are probably really slow by modern standards, and the negative rail (pnp?) output transistors are probably slower than those from the positive rail. More frequency gain compensation within the circuit, or preferably better output transistors might help.
Very interesting video as I am about to build my first transistor amp. What is that aluminium foil across fuses of JAT501? Is it supposed to protect fuses? Thank you for all your videos.
What about C3 = 47pF (upper left corner of the schematic) between the current sources ? A higher value capacitor maybe "cure" the problem or "some play" with the ratio of the two resistors of the second (for VAS) current source ? At 5:11 maybe it is a second harmonic distortion . What do you see with FFT analyser ?
car amp transistors might be cheap copy of toshiba or sanken. without verification cant tell. original toshiba 2SC5200 can survive 25V rail with open bias or shorted output without protection
You repeatly make statements about using JUNK parts and then on the JAT801 you use JUNK 20+ year old output transistor!!?? You can't evaluate this JAt801 based upon these output transistor's. Trash these transistors!!!
1: Usually is not good to use component as mechanical distance, thermal expansion make the lead break at case in long time.
2: Another bad thing that usually should be avoid is a big hole in middle of pcb , that hole will be in ground plane , and different frequency may go different path .
3: I usually use two transistor as thermal sensor
4: clipping show that transistor get saturated and the don't come out , you can put baker diode in VAS so you get headroom.
Another test to find out which one is saturated is put noise signal ( 20k limited ) and watch for top if they can come out from saturation.
5: phase margin and gain margin if you have networks analyser
6 : cut transistor lead (big one), wire it to pcb .
7: put 100p cap parallel wirh feedback resistor
I agree with this idea, especially since the output transistors are on opposite side of the PCB. Put two thermal sense transistors in your design, one on each side of the PCB near the output transistors.
love you amp board tests and as always God bless
I have been a sub for many years, and this is a great project to build onto your channel. I am keenly looking forward to seeing how this develops and hopefully, there will be an option to buy the kit sometime soon.
I love this channel
Another wonderful video
Looking forward for your successful JAT801
I got the same type of transistors! Toshiba ( 2SA1943 and 2SC5200 ) 10 of each and I really wanted to make a nice amplifier with these so I hope your project works great!
Great video as always!!!👍
Can't go wrong with that combo. I have them in 2 power amps, and they were used in most of the good brands.
The output transistors are 12MHz vs. 30MHz in the JAT501, and the cob is also about half that of the transisitors used in the 501. I don't know if that's the issue, but changing outputs in an amplifier, sometimes has the weirdest effects.
Ive heard that faster transistors are more stable if feedback is involved because it keeps the phase closer to ideal
The audio amplifier frequency is upto 20khz, so 12 vs 30 mhz will make no difference
I've been thinking along the same lines. He used old transistors not the usual sc5200 and sa1943. I'm not convinced this is a failure at all. keep pluggin John.
Maybe some compensation in the feedback network?
Chin up old chap I know you will get to the bottom of it !....cheers.
Very informative sir, great video.
I wonder if it might be useful to do some open loop tests. I think it can be hard to say for sure that it is the VAS stage when the loop is closed. Remove the feedback resistor R12 and put a very small signal in to see of the VAS stage is clipping that way without feedback. Of course you probably want to run it with very low voltage rails to keep from blowing things in open loop mode. I bet it is the output transistors being too slow that is the cause.
Perhaps an easier way to measure the open loop performance would be to short out R13 so that you were just getting DC feedback. That would be easy to do as a temporary step as you would just have to add a jumper and you wouldn't have to deal with small dc offsets on your input test signal.
could be that it needs the L1 air coil plus the 4R7 or 10R resistor in paraller for a bit of stabilty
Wouldn't hurt to try different outputs. I assume the Ft is relatively low with those old units?
Ive never seen clipping like that, ive never had stability issues with amps ive designed yet, guess I just get lucky.
I purchased a ljm20.5 amp module off ebay and it oscillates like crazy, it unusable, I had to add miller caps and a output low pass, just to get it to work, it still just slightly unstable.
Yeah, we're running a little bit hot tonight
I can barely see the road from the heat coming off it, I say
I'd be interested to see how it performed if you removed one of the output transistors off each rail. This is just academic interest. Obviously doing this would defeat the aim of creating an amplifier with higher output than the JAT501.
John, I noticed a (barely audible) high frequency when the camera is approaching the board, for example @14:46. I wonder if it is related to your oscillation problem. Sounds to me like an interference from a nearby device.
I don't like that asymmetric ringing that the 501 also has, I find these kind of issues very hard to trace because the whole amp is within a feedback loop I can't tell the difference between cause and effect, think I would blindly just go in and add 22pf ceramic caps across base, collector all transistors and if that cures it start removing them to see which transistor is the cause. I'd Like to know if there's a logical method though
I was wondering how the poles of the output might have different stability. While circuit board layout is a possibility, I think you may have hit the nail on the head. I understand it is actually physically impossible to create perfectly matched pnp and npn transistors. If memory serves, it wasn't until the mid 70s that closely matched pairs were manufactured.
@@Roger_Gadd those output transistors were from the same amp so unless he's mixed left and right pairs they are matched so I'm thinking as you say pcb layout or it's predrivers but I'm from the repair side of electronics not design so this is outside of my comfort range
@@darrenmurphy6251 I confess to being an amateur, but I enjoy electronics and learning from John's videos.
From my personal study of transistor theory, I understand that although the gains of a pnp-npn pair may be matched, one or more of the internal capacitances, resistances, and time delay of the opposing transistors won't match.
The output transistors used in this prototype are probably really slow by modern standards, and the negative rail (pnp?) output transistors are probably slower than those from the positive rail. More frequency gain compensation within the circuit, or preferably better output transistors might help.
Very interesting video as I am about to build my first transistor amp. What is that aluminium foil across fuses of JAT501? Is it supposed to protect fuses? Thank you for all your videos.
I blew the fuses during a test. I didn't have any of that size so foiled them.
What about C3 = 47pF (upper left corner of the schematic) between the current sources ? A higher value capacitor maybe "cure" the problem or "some play" with the ratio of the two resistors of the second (for VAS) current source ? At 5:11 maybe it is a second harmonic distortion . What do you see with FFT analyser ?
That capacitor is a 47uf
@@JohnAudioTech sorry John. Screen capturing is not clear enough to see the difference between "μ" and "p". What FFT shows ?
car amp transistors might be cheap copy of toshiba or sanken. without verification cant tell. original toshiba 2SC5200 can survive 25V rail with open bias or shorted output without protection
is the color package of transistors affecting the quality of it?
Lol
Small sink, no ribs, no airflow under
Change those outputs...
You repeatly make statements about using JUNK parts and then on the JAT801 you use JUNK 20+ year old output transistor!!??
You can't evaluate this JAt801 based upon these output transistor's. Trash these transistors!!!