awb50 ,a simple power amp

[infinia]

I have to use four pieces of 1N4936 diodes to get about 100mA across each emitter resisters (1N4148 & 1N4004 diodes gave too high bias current). I am still experimenting with the types & mounting of thermal-compensation diodes.

Gets very hot don't it (this high but simplistic bias scheme was my 1st concern of this design)
Try one 1N5819 in combination with three 1N4004 PN junctions should lower it alot more. Tracking is not ideal but better than most others that use 3 PNs and series resistor or plain jane multiplier. You have a larger HS so monitor the bias at your temperature extremes for checking this.

 

[infinia]

This design might work better if the drivers are mounted on the same OP heastsink, or use darlington devices instead. 1N5819 schottkey tempco is -1.1 mV/*C at about 10mA is decreasing at higher currents. see data sheet.

 

[sng001]

Hi infinia,

Thanks for the suggestion. The first channel is for proof of concepts and it was done in the typical DIY style. I just happen to have four 1N4936 in my parts bin & they are not easy to buy in Australia.

I am thinking of building the 2nd channel with one pair of ON-semi ThermalTrak device (NJL3281D/1302D). My target is only 60W/8ohm so I think one pair should be sufficient. I will mount the drivers onto the main heatsink & may try 1N5819 or other schottky diode if the bias is too hot.

Cheers, Stanley

 

[infinia]

Hi Stanley
I think this design tries for simple and supposed good thermal tracking at the expense of high bias and inefficiency. If class A bias is more agreeable then good,
Thermal track devices are costly and the diodes are really not ideal if used with random currents,
I expect your gonna pay well over $2 USD for those 2 NJL diodes and it still doesn't solve the other 2 diode mounting issues

 

[wahab]

There s some good work that has been done by there...
Unfortunately for the technical discussion, i wasn t there
starting from the end of april, as i went to my native country,
Algeria to end building a house...really a difficult project !!..

I m back since yesterday, and will provide all help needed
to the builder of this amp, who did a great work, btw..

cheers to all diyers...

 

[sng001]

AWB50 build report

I m back since yesterday

cheers to all diyers...

Welcome back Wahab,

I did built the second AWB50 board using 2SA1360/2SC3423 as VAS, 2SA1837/2SC4793 as driver and one pair of 2SA1943/2SC5200 OPT. I bought some silver mica caps so I expected that the second board should sound better than the first board.

Surprisingly I like the first board more, the treble is clearer and the bass is stronger (the first board uses two pairs of OPTs). I perceive that the first board (dual OPT) has got better resolution so I transferred the silver mica caps to the first board.

I made some changes to your design, I added a 330nF ceramic capacitor (X7R type) in parallel to C7. I used a wire-link for R4 & R9. I used a 10uF film capacitor for C1 (Input DC block). I added the tracks for a suck-out capacitor (C24) on the second board but I have not installed any capacitor in the position.

I have trouble with thermal-tracking diodes as I got only one set of 1N4936 in my parts bin and that works very well. I used one 1N5819 in combination with three 1N4148 on the second board and I experieced thermal runaway when I used a small heatsink but the fuse & the safety resisters saved the transistors.

I have build two versions of SymAsym (MikeB's ver5.3 & AAK's latest vesion with Dual OPTs), DX's BlameST & the original Leach. They are all pretty good amp & AAK's SymAsym is my favorite.

I have two CD players, one is a 25-years-old CD-304 and the other is a Pioneer BDP-51 Blu-ray player which is also a very decent CD player. I recently bought a Logitech Squeezebox3 and I will build a decent power supply for it. It seemed that your AWB-50 matches very well with the Squeezebox3 as the SB3 is a bit bassy & mellow. The AAK SymAysm matches very well with the Pioneer as the BDP-51 has got crystal-clear high.

I will build the baby-AKSA55 when I got some time.

I have attached two PDF files:
- one is the original awb50-b1-v0 (two pairs of OPTs) - this board is 4.1 inch wide.
- the second one is awb50-b1 (single pair of OPT & TO220 driver) - this board is 4.3 inch wide.

Take care with the page scaling when you print the PDF file.

- Stanley

 

[infinia]

I have trouble with thermal-tracking diodes as I got only one set of 1N4936 in my parts bin and that works very well. I used one 1N5819 in combination with three 1N4148 on the second board and I experieced thermal runaway when I used a small heatsink but the fuse & the safety resisters saved the transistors.

Hi
I'm not surprised using smaller 1N4148 diodes gives higher initial bias coupled with a smaller HS things can get away real quick. Did you ever try 4x 1n4004 compared to my recommended 3x 1n4004 + 1A schottky? I think using your 1A Si fast recovery will give higher OT bias than 1N4004, FWIW 'fast' gives no real advantage here at DC. You need a big HS for things to stabilize with-in time (or better thermal coupling ) for any hope over full operating temp.

 

[wahab]

Hi, SNG.

For thermal tracking, i also used high current diodes, in my case BY229.
In the string of diodes, i did settle the current by puting one or
more diode in paralel with the existing ones..

As you can see in the schematic, i also used the BF469/470 pair.

For the recent project, i used the 2SC2705/2SA1145 which are
the TO92 siblings of the 1360/3423 pair..

Still i have no more the original part, i can t say if the phillips
devices are better than the modern ones..

Anyway, great work , i m all ears about your audio tests..

cheers,

wahab