My latest project - fantastic power amp

[renx]

Hi everybody,
I just wanted to show you all a truly amazing power amp I have been building for the last couple of months, with very good specs...

power.........................225w/4 ohm, 160w/8 ohm
slew rate...................115V/us!!!
freq. response...........10Hz-105kHz (-3dB)
THD............................0.004% 100W RMS/8 ohm
..................................0.008% 180W RMS/8 ohm
..................................0.009% 240W RMS/4 ohm
TIM.............................0.005% 140W RMS/8 ohm (50Hz/7KHz 4:1)
s/n..............................>100dB (100W RMS)
Damping factor............>450 (150W RMS/8 ohm)

It is called "Sigma" and it has been designed by Dr. Borivoje Jagodic from Serbia. Unfortunately, his web site is mainly in serbian http://bas.elitesecurity.org/.
He is a very helpful and understanding person, also an excellent english speaker, you can contact him by mail.

Anyway here is the schematic http://bas.elitesecurity.org/sigmasema.pdf

Its amazing slew rate is almost twice as big as Brystons 4B SST (60V/us) which is twice as powerful as Sigma.
Minimum figures to meet Jung and Pass criterion for audio quality is 65V/us for an amp of 250Watt/4Ohm, 200kHz bandwidth.
As you can see Sigma is fantastic...



Here is the pic of the pcb. I consists of the amplifier, soft start circuit, +-10V PSU, line-in de-balancer and a DC protection circuit for the speaker.
It has been custom designed with copper thickness of 107um (usual pcbs have 17um thick copper).



This is a pic of the +-75V PSU pcb needed for the amp.


The 220V:57V 500VA custom made transformer (more than 4kg in weight) for the 75V PSU and a small one for the 10V PSU.



The heatsink.



The back panel made of plexi, laser cut and engraved.



A simple metal case (also custom made).


The stabilized 75V PSU. 12x4700uF, 35A rectifier...



Populated PCB. The output transistors used are matched Hitachi 2sk1058 and 2sj162 power mosfets.





Both pcbs together, as a mono block.





The two mono blocks finished driving my Dynaudio BM15 speakers.

It can be modified to have a bandwidth of 200kHz and a little nicer phase characteristic by changing the input and output capacitors, but then the stability comes to question.
In this case it makes more sense to use 192kHz converters.

I haven't done any measurements yet, since I have no admittance to the expensive equipment needed. I will post the results as soon as my university starts and the labs open.

Many thanks to Dr. Jagodic for this exceptional design, it truly sounds BEAUTIFUL...

 

[bcarso]

Quite a project!

What do you suppose is the determinant of signal-to-noise? I suspect that 100dB could be improved a good deal.

Nice to see the balanced input option.

 

[JohnRoberts]

Nice job. :thumb:

I don't notice any current limiting so be careful when wiring up speaker leads. I guess the PS rail fuses will pop if output is shorted, before anything releases smoke.

The spec you labeled as TIM, looks more like SMPTE IM. TIM is a different test (actually there's more than one kind of TIM test), where a band limited square wave is mixed in with a sine wave, but it's basically looking for slew limiting so your amp should do OK.

JR

 

[Svart]

hmm I'm not liking the toroid-on-top-of-toroid mounting..

 

[renx]

All the specs are from the designers web site, signal/noise and TIM. I will do a measurement myself and post the result.

Svart: I am convinced that most of the magnetic flux flows through the core of the transformer, since feromagnetic metal poses a much lower magnetic resistance than air. I measured the output voltages on both of these transformers when mounted like this and there was no difference. The signal is rectified and stabilized so all of the ripple from the magnetic crosstalk, if there is any, is killed by the huge capacitance.
You are right that this is not the best way to mount them, but I was limited by the space in the metal case and also no audio signal passes through these transformers. There is absolutely no hum observed, so I guess that no flux is going out of the transformers.

Thanks for all your opinions, keep them coming. :wink:

Renx.

 

[Gus]

What is the turn on surge current in the diode bridge?

Is there a slow start circuit in the AC side of the circuit?

 

[renx]

Gus: I haven't measured the the actual turn on current surge, because my equipment is too slow. But, I have simulated this situation on EWB and it showed around 90A dropping down to 6A in first 100ns. This is only for the 50000uF behind the rectifier. Of course, there is also a 500VA transformer.
I have included a soft start switch circuit because the lights flickered in my room when I turned it on and I think it is a good practice to have a stabilized power coming out of the PSU when turning on the amp.