Death of Zen 15 watt Power Amplifier

I have been watching some finished amps pictures and started to wonder a little bit about user safety. Since TO3 transistors have the collector hooked up to their case and most designs leave the side where the tr is mounted on the outer side of the case, aren't you leaving a relatively high voltage source uncovered?
Are there any ways to insulate the tr case? Or how have you fixed this problem?
 
Shaan, this looks like a nice amp but should the opamp not be decoupled ? 100 nF from + to GND, 100 nF from - to GND and a 100 nF from the + to the - pin.

Same counts for the zener diodes that could use some decoupling with 47 to 100 uF.
 
Last edited:
I asked on the local stores and they couldn't give me any good advice.

Any more thoughts? I really believe it's kind of dangerous to leave them uncovered..
I think it depends entirely on the circuit they are used in. I don't think there is any big danger in this particular amp. The voltage isn't very high.
I've seen other amplifiers that had metal covers over them. Imagine a metal U shaped piece that covers over the center of the heatsink where the transistor is mounted. There are any number of creative ways one might accomplish something like this.
 
I have been watching some finished amps pictures and started to wonder a little bit about user safety. Since TO3 transistors have the collector hooked up to their case and most designs leave the side where the tr is mounted on the outer side of the case, aren't you leaving a relatively high voltage source uncovered?
Are there any ways to insulate the tr case? Or how have you fixed this problem?

I like to see the transistors! Mines at 30v so its fine 😉

An update while I'm here. Quite amazed this thread has had over 20,000 views. My DOZ is still going strong and it gets used most days 🙂
 
Last edited:
Unfortunately, no. 5200 is a very fast bjt whereas 3055 is much slower.



Zobel network of 10ohm+100nf in series from output to ground is a must.
5200 has two problems with doz.
1- prone to oscillation without compensation.
2- Severe Iq instability resulting in overheating.
Being a high gain bjt it keeps passing more current as it heats up. DOZ is a current feedback amp and makes the best out of 3055 which is the ideal choice. 5200 won't probably be any better soundwise.


I agree, however there is an aspect of this circuit which may have escaped notice. The frequency transition of BC560 and BD139 are 150Mhz and 50Mhz min and they are relatively "fast".

"Slow" devices like TIP3055 and 2N3055 are good only for 2-3 Mhz but due to their presence which is by design the circuit has an inbuilt dominant lag in the negative feedback characteristics and it should be stable without compensation capacitors.

Substituting a 2SC5200 - a 30-50 Mhz device - changes everything as the circuit runs at greater speed and a compensation capacitor and zobel network will be necessary to keep things in check.

If an amplifier with 5200's sounds bright and overheats it is not unconditionally stable.

I would avoid 5200's and would prefer TIP35C which are "slow" but rugged.

If 5200's are a must then one compensation method would be to connect a capacitor from the collector of Q3 to the emitter of Q1 - value 330p.f. to 470 p.f. plus the zobel network mentioned above.

From observation some heatsinking arrangements could be upgraded in the interests of component longevity. 0.3 degree C rise per watt rating is suggested. Also a d.c. protection circuit would be a prudent addition.

Michael J
 
I am starting to draw the PCB for my shot at this amp!
As a starting point I was going to build the schematic provided in the westhost project 36 site and only one channel.
The power source will have a 25-0-25 120VA transformer.
My first guess is that the supply voltage on load will be around 32-33v. What voltage should the zener diode be rated for?
 
I am starting to draw the PCB for my shot at this amp!
As a starting point I was going to build the schematic provided in the westhost project 36 site and only one channel.
The power source will have a 25-0-25 120VA transformer.
My first guess is that the supply voltage on load will be around 32-33v. What voltage should the zener diode be rated for?

I would caution against using this transformer unless you regulate the supplies and reduce the amount of current drawn by the output stage.

Michael J
 
I would caution against using this transformer unless you regulate the supplies and reduce the amount of current drawn by the output stage.

Michael J

I spent the night doing some simulations on LtSpice regarding this subject.

With an input 1khz sinewave 600mV peak to peak, I needed to bias the amp with about 800mA in order to get a smooth unclipped output. That would give me an input power of about 30Watts.

Event though I'm only using half the secondary winding so the power rating should be about 60VA, I'm starting to have second thoughts about reusing this part.
 
Even though I'm only using half the secondary winding so the power rating should be about 60VA, I'm starting to have second thoughts about reusing this part.

Of course. For 30w idle dissipation per channel you will need a 200VA or higher rated transformer if it's a stereo setup. Also this will provide better supply regulation, as DOZ can demand upto 1.5times the Iq at high volume and/or signal peaks.