DSE A2760 Amplifier (aka Koda-261)

[Lectran]

I own two of these amps that have been a dependable product
that produces a workable sound for its price.

Recently both amps started burning resistors R222 and R268
which activate the dethump relays. Though some burning smell
emanated from the units there was no interruption to their functionality.

I've uprated the power rating on the resistors so far with success.

Has anyone else found this problem and know of the cause, and a cure?

Many thanks in advance,
Lectran

[Lectran]

I worked the problem by replacing each 10 ohm 5% 1 W Resistor with 4 x 39 ohm 5% 1 W Resistors in parallel.

These 4 parallel Resitors have brought the impedance back to within the original 10 ohms working tolerance and mean that each 39 ohm Resistor has only 1/4 of the power to dissipate.

Lectran

[Greg Erskine]

Hi Lectran,

These resistors are part of the zobel network. Usually if they overheat it is an indication of a problem. Are you using weird cables or speakers? 1 watt resistors are usually OK in this position.

http://users.tpg.com.au/gerskine/gre...0amplifier.htm

regards

[Lectran]

Hi Greg,

Thanks for your reply and for the link to your site.

I previously found your site through Google and very much appreciate the info and images you've provided there.

I'm driving one 8 ohm Speaker per channel in a small hall about 10 metres long. Each Speaker is fed from a figure 8 10 amp speaker cable about 15 metres long.

The Resistors in question are subject to the output voltage swing ...

the larger the swing the more power they dissipate

the smaller the swing the less power they dissipate

because each Resistor and its associated Capacitor integrate the output Voltage.

The only variable that comes to mind is that recent performances
have had a richer harmonic content meaning the amps were pushed harder for the same apparent music level.

If that be the variable then the Resistor/Capacitor integrator worked harder than previously and may simply be under-designed.

[reszipsa]

The DSE A2760 amp has mediocre performance, but can be substantially improved with a few inexpensive mods - see attached PDF file.

The amp itself is solidly constructed and can still be purchased cheaply on ebay (but not from Dick Smith - long been out of stock!)

The total cost of the mods is less than AU$20.00

I've also attached THD spectra for the stock and modded amp (LTSpice simulations). In practice, the modded amp sounds very pure - similar to a Class A amp, but runs barely warm.

I have modded two of these amps, and both have been running for more than a year without any problems.

One of the modded amps came 2nd in a recent solid state amp shootout at the NSW Audiophile Society (beating much more expensive amps).

The usual safety precautions should be taken when working on a mains powered device.

[reszipsa]

Oops! Here's the PDF file which describes the mods to this amp.

[reszipsa]

And here's the THD spectra for the stock and modded amp.

[reszipsa]

aaa

[reszipsa]

And here's the LTSpice schematic for the modded DSE A2760 amp.

Please note that:

1. The component designations in the attached schematic do not correspond with those used in the Koda schematic.

2. The following SPICE data was used for the LTSpice simulation:

2SC5200 = 2SC3281=MJL3281A
2SA1943 = 2SA1302=MJL1302A
2SA733=BC557
2SC945=BC547
2SC2383=BF446; 2N3700; 2SD1291; KT630G; ECG382; ECG31; D1812
2SA1013=ECG32; BFT19; 2SB1212; 2SA965
2SC227=KT601A; 2SC217; 2SC3420; D826; 2SB492; B492; 2SD2147
2N5551=BC546B; BF391; BC538-16; BCX49;
2N5401=BC526B; BF492; BF491; BF398; BF397
2SD669=2SC1847; 2SB649 (BD139)
2SB649=BD238; 2N6211

Incidentally, simulated THD at 50Wrms/1kHz/8 ohms is approximately 0.001% (and the stock amp has about 0.3% THD under the same conditions).

[reszipsa]

And here's the (edited) LTSpice library data file with SPICE models for all of the bipolar devices used in the LTSpice schematic for the modded amp that was posted a few days ago.

Unzip the attached "standard.zip" file, then rename "standard.txt" file to "standard.bjt". Overwrite the existing "standard.bjt" file in your LTSpice \lib\cmp folder with the renamed file.

You can then run simulations for the modded amp with LTSpice. Don't forget to turn off compression in LTSpice - click on Tools\Control Panel\Compression and uncheck "Only compress transient analyses", "Enable 1st Order Compression" and "Enable 2nd Order Compression". If this isn't done, the simulation results will be degraded.