Can some one explain me this amp?

[Tarzan]

Hi, I stumbled acros this amp and I like to understand it.
I do not have it on the nbench any more so I can't give more details as given now.
It's a commercial design and I do not wish to disclose the manufacturer.

Here is a link:
http://docs.google.com/Present?docid=ddmq8v3q_8cdhv7fcr&invite=4dc5vw
But if you like mail me and I send you a good copy in pdf format.

1. If you look at the PSU, Where is the ground?
Are Q12 and Q15 setup to create a virtual ground?
2. Q14 and Q17 must be the power stage and Q5, Q18 the drivers.

At the moment I received the amp.
PSU part:
fuse F2 was gone.
R26, R73 of the -15V burned.
AMP part:
R16, 17 and Q5 gone.

After replacing this is what I noticed:
D16 is not lit (LED is OK)
+15V is present at R44. +1.8 over D16.
+20.9V at J1 with no load connected.
0V with load.

Why is D16 not ON?
Voltage to low but why?
I checked the parts around D16 and they are all OK.
All other power supp;y voltages measured normal.


What else can be said about this design.

 

[peranders]

The link you gave us is not particulary good.

 

[Jan Dupont]

To bad quality to see any details.......
Post the PDF here.....

 

[Tarzan]

I tried to make the file smaller in size but could not.
Every body who wants to have a look can write me and I'll send the pdfs.
Unless some one can tell me how to make a pdf file smaller and how to make them available?
I tried with importing it in Corel draw and exporting it to jpg. I could go to 250K. The limit is 102.4K.
Tarzan

 

[Tarzan]

Of course my address:

VMI at Advalvas dot be

Sorry, forgot...

 

[Jan Dupont]

You got email

 

[Tarzan]

Must be in your mailbox by now.
Tarzan

 

[Jan Dupont]

Got it...
Strange how these files can be so large ??????
I'm now trying to reduce the sizes, but have failed sofare

 

[Jan Dupont]

Here is the first file

 

[Jan Dupont]

And the second one

 

[Tarzan]

Great, how did you do it?

Many thanks for the help.

Tarzan

 

[Jan Dupont]

Regarding your first question about virtual Gnd, you are right.
Left side of the schem is the Power Amp and the right side of the schem is the circuit for virtual Gnd.

Seems to me to be an expencice solution

 

[Jan Dupont]

I used the PDF optimizing function in Adobe Acrobat Pro 6.0

 

[peranders]

I just did a "save as" and the size became 40% of the original.

 

[Tarzan]

Aha, that's done with acrobat the full packet and not the common reader.
Thanks for the tips Per-Anders and ACD.
Tarzan.

 

[unclejed613]

it looks like another "we designed it this way, just because we can" design. looks like a pseudo QSC design. actually it looks like an asymmetrical bridge amp, with one side driven with bipolars, and the other with MOSFETS. i was trying to design an amp using a similar "current seesaw" as the input/VAS stage they used here, but it was a real bear to compensate, so i decided to go with a discrete op amp instead

 

[Tarzan]

It's a 500 Watt continuous power amp used in an active subwoofer system.
As the Fets Q12, 15 are used to create a virtual ground, the BJT's Q1, 17 are used to drive the speaker.

I can't understand the use of Q1, 2, 3, 6 and Q4, U1.
Is it a smpsu, but to do what?

I also see no protection for the speaker?

The left side of the schematic is a normal power stage but 500 Watts with just 2 bjt's?
Or is the virtual ground shifting towards either supply according to the input signal by means of U2 that drives both fets?
Hum, must be the way.
Thinking loud can shed light on the design.
Your comments and ideas are welcome and correct me if I'm wrong.

Data sheets:
HCPL-3100 http://www.ortodoxism.ro/datasheets/hp/HCPL-3100.pdf
2SC5200 http://www.ortodoxism.ro/datasheets/toshiba/3228.pdf
2SA1943 http://www.ortodoxism.ro/datasheets/toshiba/3176.pdf

Tarzan

 

[Jan Dupont]

According to the schematic, it's 600W into 12 Ohm load = app. 7A

It will for sure fries up if you run it with a 4 Ohm speaker at high power

 

[unclejed613]

iy functions similar to a bridged amp, but instead of two identical output stages driven out of phase, one output stage drives the speaker, and the other moves the ground reference from rail to rail. no DC protection is required because the speaker is coupled through the power supply caps.

 

[lumanauw]

Its a subwoofer amp using a "step down" switching regulator.
It's a single rail switched to have symmetrical +/-rail. For example, we only need +90V to have +/-90V if this +90V is floating and the top or bottom can be connected to GND. When the top is GND, we got -90V. When the bottom is GND, we got +90V.
The GND switchers are mosfets IRF630.

In the power supply section, after bridge rectifier it is about 140VDC (100VAC*1.4).

Point (1) and (2) is the output of the SMPS. The trigger to this smps is (3)(4). (3)(4) will conduct if the voltage is at the threshold (voltage divider of R6-R7 towards 25V). The minimal voltage for output transistors C5200 is 25V-D37.

When the output swings bigger than that, it will have a constant VCE due to this (3)(4) trigger always maintaining a constant voltage bootstrapped towards output node.

This makes the heat dissipation very little (with this supply rail modulator).