• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Update for Elektor Crescendo Power Amplifier 1982

I have lived with this 180 W amplifier for almost two decades. It has been proven to be good sounding and reliable over the years, even though I must admit a little tricky sometimes. In the posts in the www.diyaudio.com I am glad too see that the Crescendo has good reputation, and many have built it and likes it a lot. The Crescendo is still very competitive I think, and should be a contender to the Hall of Fame of diy-amplifiers, if there is one!

My guess is that almost everyone who built this amplifier has or had some problem with high frequency oscillations and perhaps some hum induced from the power supply. High frequency oscillations can be hard to get rid of. There have been some tweaks published in the Elektor I have tried, beside some that I have collected from postings on the Internet.

My amplifier had a lot of high frequency oscillations, and the RC-chain on the output went up in smoke. But the solution was easy; a small capacitor across the bias resistor solved it! I think it should been there in the first place. It is common practice in many costructions.

So, here is a listing of proven solutions to kill or reduce HF oscillations:
  • 47nF in parallel with the bias resistor P1. This killed all mine HF osc.
  • Using non-inductive source resistors R27-R30 is recommended by Elektor and have worked for others.
  • Put 1 nF between n-channel source resistors R27 and R28 is also recommended by Elektor.
  • Move the RC-chain R32 C16 from PCB to the speaker chassi posts can kill some HF osc. This is a solution used in various builds for this reason.
Suggestions for improvements of stability and performance:
  • Replace cascade transistors T8 and T10 BF469/470 with MJE350/340. This may improve sound quality. Or not.
  • Active DC regulation for lower DC offset. Could also improve sound quality.
  • Resistive collector load for the voltage amplifier (it sees a varying impedance in the gate of the mos fet's) may improve over all stability. This is used in many constructions.
  • Different feedback paths for LF and HF signal could improve stability. This is used in other constructions.

I still have some problems with hum, and I have rewired a lot a tried many other things. Nothing helps. Stabilizing the input section didn't. To compact housing perhaps? Left and right channel does not hum equally either, and it is very irritating!

On the other hand, one of the channels have a defective mos fet, but still works...

Anyway, I think this amplifier deserves another decade using some of the tweaks. I was a bit disappointed with the new Crescendo ME, I can't use it to upgrade the original Crescendo, and there are problems with high frequency oscillations in that one too!

I will be happy to receive comments and help to update the reference columns with links to diyAudio forums.
 
Last edited:
I once built the mini-crescendo version for a friend of mine (the mini was a scaled-down 70-90W version, only 1 pair of output devices and lower rail voltages).

Strange enough it would oscillate when I increased the idle current to about twice the recommended setting. I tried to run it more into class A only to find the zobel network resistor going up in smoke. Restoring it to the recommended setting solved the problem and made it run stable even into highly capacitive loads :cannotbe:

I must admit I never investigated this phenomena any further but settled for this easy way out. Probably something to do with changing mosfet parameters.

I must agree that it's a very good sounding amp.
 
The original design by Elektor had some nasty switching glitches and started oscillating in the wild. After making a new PCB(by hand drawing) for it and modding it is a super stable amp. Not a powerhouse, 100W 8Ohms at +_60V), but enough. Still use them, but every 10 years change all the electrolyt. caps./wiring.
The amp.is build in complete symmetry. Crosstalk at 20Khz= -100DB. Best I ever achieved with an 1 transformer stereo amp.

An externally hosted image should be here but it was not working when we last tested it.
 
I haven´t built the original (but the newer ME)
but have read several times that the biggest oscillation problem are caused by the output resistors (0,22) if they´re of a "high" inductive type.
Many exchanged them with for example metal-oxyde resistors.
(if you can´t get them at that power rating, just parallel 5*1Ohm 1W resistors)

greets
Jens
 
Wynand

Are you still interested to get the original article ? :confused:

I can send it for your email ;)

Ps. I have today bought a "functional" Cresendo 180W for a symbolic amount of cash compared to the real cost of the parts/project, the owner informed me was not up to the task of getting the project finalized to perfection :bawling:
The issues he had incounted was the ones bjornagain has posted the solution of in this tread ... :devilr:
Thx bjornagain :)
 
www.hifisonix.com
Joined 2003
Paid Member
I'm not surprised there are oscillation problems with this amp - where is the compensation? I read the original amplifier article where they discussed using the input capacitance of the output mosfets as part of the compensation scheme. Besides the fact that this IMO is a highly dubious engineering approach, one of the things mosefet manufacturers do is regularly shrink their mosfet die in order to improve the yield in their fabs - this also leads to changes in device characteristics - normally not enough to cause major issues, but if you have a design that's on the edge, then it will.

This amp probably has parasitic problems in the output stage and needs some gate stopper resistors (100 Ohm to 470 Ohm) and it needs some dominant pole compensation - either across the colletor load resistors on the input stage along the lines of Bob Cordell's design, or a small value cap (20-30pF) across from the collectors of the VAS cascodes to the front end diff amp load resistors. The out put inductor is also important and needs to be in the region of 2-3uH minimum with a 1-2 ohm resistor in parallel. Lets not also forget the Zoebel network. In the ME version, resistors were added from the cascode collectors to ground, presumably with tghe mistaken notion that this would set or stabilize the loop gain and prevent instability.

I think the designer of this amp tried to maximise loop gain to get the lowest possible distortion without looking at the phase marging on his amp and doing what ALL designers have to do: Make a trade off between phase margin and loop gain. The good designers get the best of both worlds, and the not so good ones end up with inadeqaute feedback or inadequate phase margin - both are unacceptable and unfortunately, the Crescendo falls into the latter category. A great pity, because with a little tradeoff, this could have been a trouble free amp, no matter who built it.

Maybe someone on the forum should model this amp and lets look at what needs to be tweaked to clean it up for once and for all.

Merry Christmas

:smash: