JAT EZ amplifier project that blows the output transistors

[@vertoboy]

Another nice video John! Few comments from me: 1)R9 and R13 are not optimized. You can calculate their values this way: R9+R13=0.5Ec/(2:5)*Ibmax (Q8) Where Ibmax(Q8) = Imax/h21(Q8) So R9+R13 = 28/(2:5)*(5/2000) = 2240 ohms to 5600 ohms Let's choose 3100 ohms for their sum then. Also R13 should be 20 to 100 times RL. For 4 ohms load it is then in the range 80 ohms to 400 ohms Let's choose 390 ohms for R13. Then R9=3100-390=~2700 ohms 2)Nice to see you are using IPS degeneration with 100 ohms resistors (R4 and R5) and this provides 5:1 degeneration ratio allowing larger differential input voltage in the linear region of the IPS. Also you can decrease the Miller cap to 68p to further increase the slew rate. However IPS LTP degeneration improves 20THD and worsens 1THD performance by reducing the IPS gain 5 times leading to 5 times lower OLG and NFB at low frequencies. To fix this I strongly recommend implementing darlington VAS (addition of one NPN and a resistor will not make the current design so much expensive, but will dramatically improve the amplifier). Note, in that case you need to increase R6 due to the increased Darlington VAS double Vbe i.e. R6 =~1300 ohms. 3) If you keep the 0.22 ohm emitter resistors then you should increase the idle bias current to around 120mA. 4) This design is prone to HF oscillations and I believe the phase margin is no more than 20 degrees. You can fix\improve it in many ways, but the most simpler here is by adding a 27p or 47p in parallel with the FB resistor R10. I already built and repaired some amplifiers with Darlingtons BDW83/84C (TO-247) and they perform really well when designed properly.

[@souvicknaskar4538]

Hello sir , it's Souvick here , 1st comment on your video. And waiting for this Jat amp EZ project, great to see you again . Thank you !

[@andymouse]

Fascinating as always, any news on the JAT 501? have you done any more tests on the PCB ? hope you are well...cheers.

[@JohnAudioTech]

I got the redesigned PCBs in but have yet to populate it. I need to get on it. So much to do...

[@beamer.electronics]

Interesting video, thank you. When I experimented with this type of amp, I found C4? - the 100uF being electrolytic behaved oddly. It was occasionally spikey and random on its leakage, I changed it for a non-polar/non-electrolytic. The other issue I had was the bias pot; if just 2 pins were wired - the center contact occasionally lifted (even for a fraction of a second), sending the current to max! Using all 3 pins solved that issue. I'm also not a great fan of Darlingtons :) All the best, Beamer.

[@イエンスヨハンセン]

I always learn a little something from your videos. Thank you!

[@bernardmarx6705]

It's not about SOA only, the problem is thermal runaway, you could try higher emitter resistors for output transistors for better thermal compensation. Anyway, one pair of output darlingtons in TO220 in this case is not realistic.

[@JohnAudioTech]

Yes, I want to try different emitter resistors (I need to order some). I'd also recommend the supply voltage max at +/- 25 V instead of +/- 28.

[@christianhermann5306]

first ...very nice work John...i like your channel....easy for learning....i checked some Darlington and some BJT with "bigger" hfe. maybe a compromise is the BD909 orBu931T. you are right to check the SOA at 30V - so you get about 3 or 4Amps here to have some head room

[@paultyson5145]

Thanks for the vid. Very interesting. I really like the signal tracer gadget... never seen one of those. I wonder how they work?? Please excuse my ignorance on that!! I will look at your other amp vids too. Thanks again

[@PeterMilanovski]

Just goes to show how good a tool like a signal tracer is! Some transformer oil in the dummy load wouldn't hurt either! It would help transfer some of the heat out and help keep your home warm during those cold winter days while you listen to some music LoL...

[@DeadKoby]

I've fixed a number of Fender guitar amps that had Darlington output transistors.

[@netrapalsinghprjapati1935]

Nice project sir👍

[@gkdresden]

John, I would like to recommend the use of Sziklai pairs (BC640/2N6488, BC639/2N6491) instead of Darlington pairs. I am sure that it is a problem with thermal runaway, which can be avoided by use of Sziklai pairs. For Sziklai pairs the temperature dependence of the VBE of the power transistor doesn't matter anymore. So you have to couple the bias spreader transistor thermally with one or both of the driver transistors on the PCB (if even necessary).

[@socratesaresto8312]

Any idea for a transistor amp for small signal (radio signal level audio) to a quarter watt up to one watt max speaker? I tried tl072 for small signal to drive an output transistor but i didn't like the distortion so i had it as a buffer while the one of the opamps did voltage amplification, still not a very good response as input level of radio reciever audio starts around 5 ma, any thoughts ..?

[@RandomUser6947]

If you like playing more with darlingtons on this amp, you could try the BDV64BG/BDV65BG transistors. They do cost a little more, but I think the SOA is about 5 Amps at 30 Volts. I think they cost about $2 or $3 a piece. They also have a bigger case and should have better heat transfer.

[@christianhermann5306]

hfe is here about 1000.....yes thats right. the soa 30V/5A is "safe" but maybe try some darlingtons with "less" hfe or normal bjt with "bigger" hfe to avoid oscillation. e.g. BD909 or BU931T

[@brandongaming5284]

Where did you get the Transistors from if there from eBay do a low current breakdown voltage test i use a Tv Led tester. I bought a whole lot from eBay had to toss hundreds of Dollars in the Bin. They where rated Collector Emitter Voltage of 250 volt but when i tested them they broke down at 60 volts.

[@JohnAudioTech]

I got them from DigiKey. I don't trust eBay.

[@DAVIDGREGORYKERR]

What about an Amplifier with 2xBD131 power transistors with the bottom transistor driven via a BC212L search for 5W Class A Amplifier which is the type of Amplifier that was used to drive the color guns of a CRT.

[@dexter-lab-42]

hello john big fan of your channel, love the JAT content! but was wondering if I could email you with some amp circuit questions

[@iblesbosuok]

Why don't you try to use power MOSFETs for final transistors. A little bit sacrifice on output voltage swing. They need more simple bias servo network. I think IRF530//IRF9530 good for start.

[@JohnAudioTech]

@electronika In his book, Bob Cordell shows that you can design a nice MOSFET amp. This is something I'll try at some point.

[@Wolfgang3418]

There are some MOSFETS especially made for switching like the logic level ones, but the aforementioned IRF-types are good for audio. A very simple MOSFET amp you find here, not really hifi, but OK kzbin.info/www/bejne/bqK5gKqNi8Sgjas

[@BlodaBlodaBloda]

And if that is just a spreader and not the full sink, then why are you trying to test your amp without the full sink attached?

[@50Hz]

Hey, any chance you could give me some advice on finding an amp that could run variable loads from 0.5W to 50W. I’m trying to power AC electrical loads but it’s hard matching a amp , transformer and load

[@stopthebus]

My first thought was a VFD - but then I re-read your question. Can you explain in more detail what you want to do?

[@50Hz]

@@stopthebus i sort of solved my problem, I'm building a AC test set that can produce AC currents and voltages. Currents is easy but voltage I had trouble with it being stable at different loads. But I did a work around by putting a loading resistor in.

[@Wolfgang3418]

This wattage deserves a bigger case, I think, TO220 is all too small. Maybe your first transistors died of overheating because sometimes the transistors lift a bit when screwed on. So firstly you could use seperate heatsinks for the two transistors without insulation tabs and secondly you could srew an additional piece of aluminum to the front of the things. Good luck!

[@gn_electronics]

I have built an amp using the bdw83c/bdw84c pair. So far it works well but it's hard to adjust the bias without the risk of thermal runaway.

[@JohnAudioTech]

what emitter resistor value did you use?

[@gn_electronics]

@@JohnAudioTech 0.27 ohm

[@JohnAudioTech]

@@gn_electronics You might try a higher value for better thermal stability, in the .39 - .56 range.

[@copernicofelinis]

This is how humans learn. Cats, on the other paw, do not commit errors in the learning process.

[@zapruderpedro]

anyone know the song John was playing? sounds like Mountain maybe?

[@zapruderpedro]

Mountain - Never In My Life

[@threeMetreJim]

Maybe darlingtons are too sensitive when it comes to bias current and heating, compared to lower gain transistors. I've never seen an amplifier using darlingtons on the output, even though output stages (AB) usually resemble darlingtons. The difference being the driver transistors aren't in as close thermal contact as they would be if in the same package. I'd put the circuit on a current limited supply, and see if the current started to 'run away' once it heats up. Edit: apologies, should have watched the entire video

[@1pcfred]

That EEVblog meter is a Bobby dazzler.

[@JohnAudioTech]

As long as I don't come a gutser as Dave says.

[@1pcfred]

@@JohnAudioTech no wockers.

[@johncoops6897]

@@1pcfred - It's spelt "wuckers". FYI: It's the shortened version of _No Wucking Forries_ .

[@VicariousAdventurer]

Design Goals: Magic Smoke!

[@watup110875]

Could have gotten a bad batch of transistor that caused that, I know 23 years ago now, my ex-daddy in law god rest his soul, bought 9 sets yes 9 sets of spark plugs for tune up and the 10 set that was bought was perfect, sad how they push junk to the public with no liabilities! This would be no different than the spark plugs.

[@kittony71]

check on bias

[@BlodaBlodaBloda]

Increase the values of R17 and R18 to .47 ohm and your problem will go away. Your heat sink is too small too.

[@Theineluctable_SOME_CANT]

A 100 nF cap across the dropping diodes will help.