p68 trouble .....

[sakis]

i constructing a prototype of this amplifier

the outputs i use is mj 15003 -4 and here is the problem :

i made a prototype with 4 output transistors that works perfect ...so i made another one with 6 out per channel ..... the second prototype also works but when testing with 1 khz sinewave and 8 ohms resistive dummy load the lower part of wave form only gets distorted in some levels....somewhere in the midle when you feed the amp with 900mv or so

after that clears out and also before that looks almost clear ....

any opinions about that ????

[kalmara]

So, my question in the P3A thread has been answered ... seems that maybe the miller cap in the negative voltage driver is just for that ... what value do you have there ? You won't loose anything if you try and tweak it a little bit? (unless you make it too low and get oscillations)
My guess is that this is some kind of disadvantage of the bootstrap ... maybe, maybe not?

P.S. Just an off-topic question - if your output devices are TO-3, then what kind of drivers are you using ? (Q5,6,7,8)

[EWorkshop1708]

Originally posted by kalmara
So, my question in the P3A thread has been answered ... seems that maybe the miller cap in the negative voltage driver is just for that ... what value do you have there ? You won't loose anything if you try and tweak it a little bit? (unless you make it too low and get oscillations)
My guess is that this is some kind of disadvantage of the bootstrap ... maybe, maybe not?

P.S. Just an off-topic question - if your output devices are TO-3, then what kind of drivers are you using ? (Q5,6,7,8)

Probably with 6 pairs he need more miller capacitance. Try 500pf. Also some 10-100 ohm resistors in the drivers bases help too.

Also I'd lower R13 & 14 to 100 ohms or less. It adds stability at the expense of gain.

Now I'm always a fan of miller caps for the drivers. I use them for both upper and lower drivers with 100 ohm base resistors on an amp with 5 pairs. It is harder to get an amp stable with lots of parallel outputs, so you have to do extra stuff to get it to behave.

But for some reason some people here don't agree with me, but it works great. Miller caps for the drivers is a great thing IMO.

[jaycee]

The lower half is oscillating. My bet is that it is because Q7 and Q8 are too slow - MJ15003/4 are 2MHz, and the MJL4281/4302 are 30MHz.

I would add 10 ohm resistors in the bases of all of the output transistors as recommended, but tbh I dont think this driver stage is going to work very well with slow transistors. You might be best simply using MJE15032/33 alone.

[EWorkshop1708]

Originally posted by jaycee
The lower half is oscillating. My bet is that it is because Q7 and Q8 are too slow - MJ15003/4 are 2MHz, and the MJL4281/4302 are 30MHz.

I would add 10 ohm resistors in the bases of all of the output transistors as recommended, but tbh I dont think this driver stage is going to work very well with slow transistors. You might be best simply using MJE15032/33 alone.

I've done it before with no problems. I've used 30mhz drivers on slower 2n3055/mj2955 Also, those MJE15032 are also 30mhz.

A miller cap on those 30mhz drivers will slow them down a bit.

[Andy L. Francis]

The MJL's Ft is ~35MHz, while the MJ's are ~2MHz.
I'm almost sure your OP stage is oscillating.

First I'd reduce R13/R14 to ~33Ohm, an see what happens.

Anyway I'd replace the diode string with a Vbe multiplier for proper bias current adjustment.
What's you current bias?

Then MJE15033 is not a good choice for VAS purpose.
Same for Q1-Q2-Q3, tthese MPSA's are slow devices with low gain.

On the margin the whole amplifier needs a redsising IMHO.
I wondering why there are many fans of the ESP amplifiers...

[wg_ski]

isn't slow vs fast outputs, or even slow vs. fast drivers. You can't just arbitrarily load down a CFP driver stage and just expect it to work. They don't scale up directly. They're finicky and layout-dependent. It can be made to work eventually, but while you all fool around with it inserting 100 different capacitors in as many locations, I'd just hook it up as a triple darlington with 1 or 2 ohm base stoppers on each output and have it working while everyone is still scratching their head.

[sakis]

a few things ......

A)most esp designs i ve made or tested work like hell

B) the amplifier is working and its a tested design by rod but let us not forget that this is a test version that i ve made on my own pcb

C)the version i made with 4 transistors also works perfectly

D) rod has a version with aditional 4 transistors and no other changes at all .....so its either my pcb dont work properly or the choise of using the circuit as is together with mj15003-4 outs is wrong ....

i will try though a bigger miller cap plus base resistors to see what happens and let you know .....

thanks for the input

[sakis]

Originally posted by wg_ski
isn't slow vs fast outputs, or even slow vs. fast drivers. You can't just arbitrarily load down a CFP driver stage and just expect it to work. They don't scale up directly. They're finicky and layout-dependent. It can be made to work eventually, but while you all fool around with it inserting 100 different capacitors in as many locations, I'd just hook it up as a triple darlington with 1 or 2 ohm base stoppers on each output and have it working while everyone is still scratching their head.

i dont exactly understand where these resistors are to be placed ......if you ll be kind enough to show me i would be very thankfull

thank you

[sakis]

values of 330pf and then 500 pf of miller cap in the vas stage made some improovement but nothing really special ......

base resistors will follow