[Hybrid Amp] Questions about the use of zener diodes

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Administrator
Joined 2004
Paid Member
Hi Johnix,
Man, is that ever close to a Counterpoint SA-100!

The relay prevents bias current until the tube stabilises, a "thump" reducer.
The BC 546 / 556 form current sources. I did the same mod to the SA-100. The zeners form a voltage reference to the bias supplies to isolate the bias voltage from supply voltage hum and variations. They are also needed to set the zero voltage.
This thing is so close to the Counterpoint, slightly improved and worse, I'd have to call it a rip off. I think I will call it a copy. The only major difference is no feedback and therefore, one less stage, and the lack of a gate drive buffer.
If I want to be kind, I'll say the design borrows very heavily from the Counterpoint SA-100. I mean heavily.
-Chris
 
anatech said:
Hi Johnix,
Man, is that ever close to a Counterpoint SA-100!

The relay prevents bias current until the tube stabilises, a "thump" reducer.
The BC 546 / 556 form current sources. I did the same mod to the SA-100. The zeners form a voltage reference to the bias supplies to isolate the bias voltage from supply voltage hum and variations. They are also needed to set the zero voltage.
This thing is so close to the Counterpoint, slightly improved and worse, I'd have to call it a rip off. I think I will call it a copy. The only major difference is no feedback and therefore, one less stage, and the lack of a gate drive buffer.
If I want to be kind, I'll say the design borrows very heavily from the Counterpoint SA-100. I mean heavily.
-Chris


Almost SA-100. The valve stage is simplyfied, and avoid the feedback. The idea almost same...

sajti
 
anatech said:
Hi Johnix,
Man, is that ever close to a Counterpoint SA-100!

The relay prevents bias current until the tube stabilises, a "thump" reducer.
The BC 546 / 556 form current sources. I did the same mod to the SA-100. The zeners form a voltage reference to the bias supplies to isolate the bias voltage from supply voltage hum and variations. They are also needed to set the zero voltage.
This thing is so close to the Counterpoint, slightly improved and worse, I'd have to call it a rip off. I think I will call it a copy. The only major difference is no feedback and therefore, one less stage, and the lack of a gate drive buffer.
If I want to be kind, I'll say the design borrows very heavily from the Counterpoint SA-100. I mean heavily.
-Chris

Correct me if i'm wrong but the voltage lvl at the node where each zener is connected to the BC 546/556 is +/-12V ?

But is there any current passing through the zeners ? Are they a kind of load for the current sources ?

Then I was wondering how many mA were taken from the current sources to the both resistive networks with the pots ?


And lastly, could this amp work with the relays off all the time ?

I still dont catch the point in having a filter network (1uF / 100Ohm) to work well ?


:confused:
 
Administrator
Joined 2004
Paid Member
Hi Johnix,
The 100 ohm resistor and cap is an attempt to decouple the tube supply from variations caused by current draw in the output. I designed a sand regulator for the Counterpoint. It makes a huge difference in a positive way.

The amp will not work with the relay removed as the DC offset created as the tube starts to conduct will blow the fuses, or the woofer voice coil. You really, really need this relay.

Yes, the supplies are +/- 12V. Most of the current flows through the zeners to ground. The current sources replace a resistor and go a long way to isolate the circuit from supply variations. The SA-100 uses 30 to 36 V zeners. The other current flows can be figured out with ohms law and some logic. Most of the current should be flowing through the zeners.
-Chris
 
Administrator
Joined 2004
Paid Member
Hey lumanauw,
The Counterpoint SA-100 sounds awful. Really not good. It has so many design faults it isn't funny... but they sold a whole bunch of them.

After correcting the biggest design flaws, you end up with way too much gain. Then it starts motorboating. It uses feedback sort of. But the power supply is problematic.

To make a long story short, after much circuit redesign, addition of PCB's, they start sounding okay. I'm still working on bias tracking issues. The amp wants to run just over 50 C.
-Chris
 
My experience is that mosfets sound good only in class A without feedback. One of amplifier I use IRF530 SE class A driven by two half of ECC82. Voltage gain about 10, and the bias for the output device is about 3.5A. Of course the supply is low, only +/-10V, so the output is 7.5W to 8ohms load.
For this output power 1-1.5A bias is enough, but I found that more bias sounds better. Unfortunately the TO-220 package is not able to dissipate too much power. So I tried high power mosfets, such as IRFP250/260. They can handle up to 80-100W, but sound worse. So, generally high current devices are not good for SE mosfet application...

sajti
 
Administrator
Joined 2004
Paid Member
Hi lumanauw,
I was going to send you something. Do you have a way to do this? 'Tis SA-100 related. E-mail me directly as it's too large to post.

-Chris

sajti,
My experience differs from yours here. Mosfets tend to sound better with a small amount of feedback in class AB for higher power than you've indicated. The larger FETs have more of a gate charge issue. I think that's what you have noticed. The gate drive circuit must be able to handle current spikes. The entire amp has to be fast due to the nature of the gate charge characteristic. P channel devices are a little different on top of that.
 
I tried some class AB mosfet amplifier, without feedback. They sound terrible for my ears. The biggest problem was the bass. It was not well controlled... With 20dB of feedback, the sound was much better. With pure ss solution i used high current VAS to drive the gate capacity. There was amplifier with 100mA biased MJE15030/31 VAS stage to drive two pairs of IRFP240/9240. The sound was good, but not as good as I expected.
The valve driven SE mosfet amp sounds great, with no feedback, and very cheap devices. Only the limited output power was the problem. Right now I'm working on the solution: Bridged version, with about 20W output

sajti
 
Administrator
Joined 2004
Paid Member
Hi sajti,
I can well imagine a class AB mosfet would sound bad. Certainly not worthwhile. You pretty much need feedback and a robust driver stage for these. I've heard class AB mosfets that sound really good and some that are pretty bad. Same for bipolar, so it's hard to judge. I don't care for the driving characteristics of mosfets, that's for sure.

How does the SE version sound with a little feedback?
-Chris
 
Unfortunately I can't get really good quality mosfets in my country. IRFxxxs are not planned for audio application, and I think they not provide good quality in class AB application. I found lot of difference between the N and P channel devices. So right now I plan BJT amplifiers, with low feedback, and class A mosfet followers with valve drive.

I didn't try to use feedback in my hybrid design. It contains two stages only: 1st is the common cathode valve with E82CC, and the 2nd is the class A mosfet follower. The open loop gain is about 10, so I have no gain to reduce. Now I think to apply two stages with feedback to gain of 1, and drive them with long tailed pair, to get double power with same rail voltages, and bias current....

sajti
 
Administrator
Joined 2004
Paid Member
Hi sajti,
IRFXXXX are used in Canada and the States for audio amplifiers. You are right that they are not designed for this specifically. Adcom uses these as well as others.
When you have enough gain to apply a small amount of feedback, I'd be interested in what you find.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.