Text preview for : P400Tips.txt part of usilvatel za manqcy-400W



Back to : p400.zip | Home

Huh, beleive or not I received more than 200 letters with questions for this amp. I have no
spare time and I'm working as a software engineer so I decided to remove project from site very soon.
I answered only a few letters and this document is actually part of one mail I sent as reply.






First of all excuse me for my bad English. It's not my native language.

A little bit of history
___________________________

I built different kinds of amps (tubes, fet, integrated stk and so).

Huh I designed a lot of different audio, microprocessor, ... circuits, but during 1999
NATO bombed us.They hit my house and all of this projects which I intend to put on net were burned.

I designed this schematic and I built first in 1994 during my studies, and I built
more than 20 amps of this kind, two of them (and one Quad 405) I still use as
Dolbby 5.1 sistem connected to my comp. My idea was to build cheap but High Fidelity,
highpower,high performance sistem. I can freely say that it has very good sound, better
than most sistems I heard, and I heard a lot of amps. Only Quad 405 (very old clasic but
still one of the best) is better in midrange.
Schematic you saw is the last variant of this kind of amps. The last one I built in 1996,
so I'll try to help you as much as I can because it was a long time ago.
( I used Textronix TDA220 2x200MHz digital osciloscope during tests.)


Answers to most common questions
___________________________

Well.... I designed circuit board for several diferrent amps. With small
component modification you can make pnp-npn or pnp-pnp or npn-npn output
stage amp, using the same board. There are no R10,R11,R14,R15 resistors,
in this application, you should use shortwire (1mm copper wire=0 ohms).
Q1-Q2 and Q3-Q4 are dif amps with common emmitors as you probbably noticed.
I used this double dif circuit because it is fast, it has low harmonic
distorsion but it has enough power for next amp stage.

RSTAB and R10 in protection circuit are also shortwire. Schematic and circuit
board for protection are different. There is transistor voltage stabilisator
on schematic but I actualy used integrated positive 24V/1A on circuit board.
I didnt' designed this schematic, I only designed circuit board. I find schematic
somewhere, but it's great. It has 2sec timer during powering up to supress BUMMMM
in speakers when you turn on amp, and it turns off speakers if there is any
DC voltage on output of amp (if anything is wrong). I like it. I used two colors,
three pin led for indication, connected on reley and it looks cool.


Part two
___________________________
There are a lot of tip&tricks for amps making.
First of all : input stage. Please use 2k2 resistors instead of 100 ohms for
R4,R5, because you can hardly find so qualitet transistors with same hfe for
output stage and rest. R6-R4-R5-R7 resistors and R3 trimer are used for tuning
output offset with no input signal which should be 0V. (Use QUALITET TRIMERS for
R3 and R21 if you want long amp life!!!) When you complete your amp you should
adust 0V on output with no input signal using R3.
I used this passive resistor network because it has minimal distorsion and it has
no influence on frequent caracteristic of amp so amp has very wide frequent range.
Try to find Q1,Q2,Q3,Q4 with same hfe, but it will work anyway.
C6,R23,R24,C5 are parts of negativ feetback.
Q5,Q7 are common emmiter stages.
Q6,R21 are used for adjusting DC current of output stage with no input signal.
It should be 30-80mA through +Vcc terminal. Put ampermeter serial with DC power
supply and adjust current with R21. My idea was to put Q6 on heatsink of output
stage to compensate current when output transistors are hot (negative temp. feetback),
but it's not necessary. Actually I didnt' do this on any of amps I made, but teoreticaly
Q6 should be termo coupled with output stage.
Use small onboard heatsinks for Q10,Q11.
R33, R34, R39, R40 are vertical mounted, for better cooling. Q8,Q9 are protection transistors
limiting maximal current to 10A. And so on and so on....


Important part
___________________________
I had a lot of troubles..... I hardly find good output transistors. Everyone makes transistors
today (Malesia, Mexico even China). It's important that output transistors have Vcc>140V!!!
If this is not case BUMMMMMM.... and transistors are burned in sec. When you apply input
signal more than 2V you teoreticaly have Vcc+Vcc~130V on Emitor-collector junction on all of
output transistors so it is very important that you have good transistors.

Output transistors characteristics:

MJ15003 NPN NF/S-L>2MHz Ucc=140V Ic=20A P=250W
MJ15004 PNP NF/S-L>2MHz Ucc=140V Ic=20A P=250W


Well this are characheristics from catalog but I found that most transistors cant' stand more than 110V!!!!!
As I said: Everyone makes transistors today. I found a few good transistors and made one amp with Vcc=80V
(2x80=160V!!!!) with big 1600W power transformator and 50mF electrolits in powersuplly and you can imagine
this monster!!!!! But my advice for you is to use 2x40V-2x60V(maximum)DC power supply and you'll still have
so much power. Amp is designed so it can be powered from 2x16V-2x120V DC power supply (well 120V if you find
transistors for output stage which have Vcc>240V). So DC voltage of power supply is not critical. Be sure to
use good power supply (massive transform. with a lot of amps, minimum 4A per chanal).
Use good heatsinks (there is a lot of power and there is a lot of heat also. I use 5.25inches cooler and
two massive heatsinks. See pics on my website. One more trick. Be sure to have good wires from emitors and
collectors of output transistors to circuit board. You also have to solder wire to cases (collectors) of output
stage transistors. Bad collector contacts (although instrument show 0 ohms ) can cause heating, disfunction or burning of
transistor. Boy we are talking about almost a half of KW sinus power!!!! There is high current for that power.
Dont' power up amp till output transistors are not on heatsink.


The most important trick
___________________________
You make amp. Ok. You should try it. Ok let see how it works. Stop!!!!!!!!! It's not 20 or 100W amp. You should
test it first.
______________________________________________________________________________
PUT TWO 10 OHMS (4w OR LESS) RESISTORS IN SERIAL Vcc+ AND Vcc- DC POWER SUPPLY.
-------------------------------------------------------------------------------
You can put it instead of fuses. It will limit current, and in worst case on of resistors will burn (if you made mistake assembling amp)
protecting transistors and rest,during tests. See if you have <80mA DC current from DC powersupply.
Adjust it with R21. Adjust 0V on output with R3.
Still dont' apply input signal. Let amp work with no signal for a 10min. Check heatsinks, which should be cold.
Amp shouldnt' oscilate. Once I have this problem but not with this amp. Everything looks fine, 0V, no sound in speaker
but heatsink was hot. I measured 0V DC and 0V AC on output but.... when I check with oscilkoscope I saw that amp was
oscilating on 2MHz.Huh.... It was bad condensator.
Ok you should connect speaker if everything is fine. You shouldnt' hear anything. No clicks, no his, or anything else.
Ok. Disconnect speaker, connect voltmeter, apply input signal and see if you have output signal.

Connect speaker apply music signal to input, slowly increase it, you should hear it without distorsion through speaker.
Let it work for some time this way, but dont' push it or 10ohms resistors will burn.


(This part is not nessesarry: Connect osciloscope, check output signal with 1KHz 0.5Vpp input sinus signal.
If there is crossovers you should increase DC current adjusting R21).

If everything is fine you can remove 10 ohms resistors, replace them with fuses. Turn on amp. Leave it for 10 min with
voltmeter on output and ampermeter in serial with one fuse, connected. Readjust 0V dc on output and DC current.

Connect speaker turn on amp listen it for 30min but keep checking heatsink temp. And if everything is fine that's it!!!!


Srdjan Markovic