ZEN/SOZ hybrid

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've been thinking about the Pass DIY amp designs a lot lately and reading old threads about peoples efforts to taylor one of Nelson's designs to suit their personal tastes/needs. I was very curious to try the SOZ but didn't for 2 reasons

1)like so many others I could not justify the gross inefficiencies of the design.

2)No matter how hard I tried I could not get my head around how the circuit works (ie. how it can be scallable without changing resistor values) - Ohms law certainly doesn't explain things. As I believe everything I do should be a learning experience, building someone elses design without understanding how it worked did not appeal.

This is what I did:

I basically used 2 ZEN amplifiers per channel in a differential pair arrangement (as in the SOZ design). The amp uses a single positive supply rail (24V in this case). The gates are biased at 4 volts via voltage dividers directly from the supply (ie. no feedback). Because the inputs are above ground potential coupling capacitors are required as in the ZEN design. The mosfets are biased at 1.5A each through an 8 Ohm resistor per side giving 12V at the drain of each mosfet. This equates to 4W RMS per channel with an idling current of more than 4 times the maximum required by the load (8 Ohm).

I've only just finished this amp so I'll reserve my judgement on how it sounds (compared to the ZEN which I built) for a few weeks.

I would appreciate input from anyone who has an opinion on any of this especially the following.

1) Resistors Vs. Current sources
2) Benefits of operating at higher idling currents/employing deeper suppliy rails (I posted this question a little while ago but nobody was interested, figure I'll try it again)
3) There was mention in a therad some time ago (I think it was grey) of the possibility of doing away with the input coupling cap. Something about the preamp of a tube amp. Cananyone eleaborate on the possibilities of removing this capacitior while keeping a positive bias at the gate.


I'm short on time, but you've tapped into a topic that is of current interest to me. (Okay, okay...along with about 20 others.)
1) I think the idea of differential Zen amps should be followed up. Note that if you can get the circuit running properly (shouldn't be hard), you should be able to remove the output coupling caps, as the DC from one amp should come pretty close to balancing the DC from the other. Yes, you'd need to float the output from ground or run one or both sides to ground through a resistor--but you'll end up floating the output anyway, yes?
2) Once I get the second pair of Aleph 2s up and running, I have several experiments in mind. One is to run Alephs as a differential pair as an intermediate step towards an X/Aleph hybrid circuit. Since the Aleph and the Zen circuits are cousins, you and I may be able to compare notes.
More later, hopefully, as I've been cogitating on this for a while...


The amp is operational and I never included the DC blocking caps on the outputs as the DC offset is very low (I verified this before hooking up to my good speakers). I haven't referenced the outputs to ground through 100 Ohm resistors as in the ZEN schematic. I didn't feel they were necessary as I have not seen any differential style amp stages using them. Is there a reason I should be using these?

I probably should note that I am driving the amp from an unbalanced input. This is not how I want to use this amp so I will be building something similar to the BOSOZ when time/money permits.

I've been trying to sneak a few minutes to post for the last couple of hours. Let's see how far I can get.
1) I'm on record as preferring resistors to current sources sonically. However, there are times when a current source is the best solution to the problem. I used one to drive the input differential to my tube amp, as it gives near perfect results when used as a phase splitter--something you could consider for conversion of unbalanced signals to balanced. Others will disagree (what? you're surprised?). Jam feels that current sources (and their sundry variations) can be tamed. My rejoinder is that lions can be tamed, too, but that doesn't mean that I want one as a pet...
2) All things being equal (as remote a chance as getting fat women to quit wearing spandex), higher current is a Good Thing. Rails can then be expanded up to the point where the device is up against thermal limits. How hard you want to push your gain devices is up to you.
3) There are two pretty good options for getting rid of those pesky input caps: First, go to a bipolar power supply for the Zen (but by the time you do this, you're within spitting distance of building an Aleph, might as well go whole hog); second, go to a depletion mode gain device. Okay...what's a depletion mode device? A depletion mode device is a tube or JFET; a device whose grid/gate is more negative (for JFETs we're assuming N-channel, which is the most common sort of JFET, anyway; tubes are sorta N-channel by default) than the cathode/source. Depletion mode devices are self-biasing, which means that you can tie the cathode/source to ground with a resistor (or current source) and ground the input (yippee! no cap!). The quiescent current flows through the resistor, giving you a positive voltage offset for the cathode/source, leaving the grid/gate negative by comparison.
If I understood your description of your circuit, you don't need to ground reference your output. It's just an option.
Running an unbalanced input into a balanced circuit requires a phase splitter (even if it's just handled by the circuit itself). If you've got one Zen input grounded and are putting the signal into the other input, you're probably not getting very symmetrical drive right now, so the circuit may now be functioning at its full potential yet.
This is probably pretty disjointed, as I keep getting interrupted by the real world (boring stuff like earning a paycheck). Ask questions as necessary.
Back to you.


Although I have no reason to prefer resistors (ie. I've not had enough time playing with these things and listening to the results to have a valid opinion about whats better) I prefer to use them instead of constant current sources due to the simplicity of the circuit. I like things as simple as possible, I think at this stage of my hobby I require said simplicity so I can see the woods without the trees clouding my vision.

Although I am using relatively high current in this circuit, due to the low power (4W/ch) the devices are well within their temperature limits. This reminds me of a question I had; If both legs of a differential amplifier are biased at say 1.5A, does the load see 3A?

Funny you should mention JFETS as I have been doing a little research into using these devices myself. Once again I refuse to construct a circuit I do not atleast partially understand so the use of these devices in this style of circuit will have to wait until I have the time to research it fully.

One thing about JFETS, how does their much lower transconductance affect their use in a circuit such as the one we are discussing? What mods have to be made?

You mentioned in a previous post using split-unequal rails to bias the SOZ circuit to increase the efficiency. I was thinking of removing the bottom 8 Ohm resistors from the SOZ circuit and with a unequal supply (eg. +20V & -5V) bias the mosfet through the 8 Ohm resistors up top and the 1 Ohm resistors down below. What are your thoughts?

Your right I am grounding one of the inputs which I understand creates an increase in distortion from balanced inputs, but at this stage I don't see what other option I have. Any ideas?


Is it possible to post your schematic or e-mail it to us. It sounds interesting.

Despite Grey's and my differences on current sources there are probably advantages to both approaches. Experiment and choose the one that sounds the best. I had suggested to Grey to try an experiment and replace the current source of an Aleph differential with a resistor attached to a higher B+ rail which is an experiment I should try it but alas I don't have Alephs anymore, but I am open to the possibility that this might be a viable solution.

After all if we don't explore different solutions to problems we don't advance our knowledge.


P.S. Depletion mode N-channel mosfets are available but they are rare and I have not been able to locate any power or P-channel devices.
I've been listening to this amp for a while now and I think I can give some observations/listening impressions. I will not make outright statements rather comparisons to the ZEN amp which I already built.

The music is detailed, maybe a little more than the ZEN. The music seems very forward almost to the point of being too bright on some tracks. The warmth of the reproduction (artificial or otherwise) of the ZEN which I so like is not absent but a little less with this amp.

I am still not sure which I prefer and for reasons stated below I will not make a judgement just yet.

There are a few things which I should point out which may have produced the characteristics I am describing.

1) In my ZEN I bypassed both the input and output coupling caps with metalised film polyprops. Although the output coupling caps aren't required I still use a 47uF electrolytic on the input which at this point is not bypassed. I will be doing this today and reporting any differences I hear.

2) As stated in a previous post I am still using the feedback circuit from the ZEN. This is another modification I hope to make this weekend.

3) I am using an unbalanced input and grounding the other input.

Could any of the above give rise to the differences in sound I have discerned. Is what I discribed above what a SOZ is supposed to sound like?

Resistors v Constant Current sources

In this particular application the use of constant current sources will dramatically lower the level of second harmonics. However it does lead to greater odd order harmonics which extend to quite high multiples. many people feel that some second order harmonics provides a "pleasant, valve like" sound. On the other hand odd harmonics are generally considered undesireable. Once again just looking a raw specs can be misleading as the THD with constant current sources is much lower, notwithstanding the fact that many prefer a design with resistors.

Hope this helps.



I have never modelled the zen amp so was not comparing different typologies, simply answering your question regarding resistor v. current sources in an op amp. I will be home in a few days and can look at the Zen if you wish to pursue that.

Regarding second harmonic v. higher order, yes I would expect a subjective comparison to suggest a greater level of second harmonic produces a "warm" sound. Depending on their amplitude and multiple, higher order harmonics can produce sound that is described as "harsh", or "bright". If not excessive is sometimes confused as being "more detailed" even though the latter case is not actually true. Describing amps in this manner can at times be a triffle confusing given the symantics involved.


Have a look at Fourier Analysis. Circuit Maker does not, as I understand it, provide as much information as is possible using the XSPICE engine, however is possible to model what we are refering to here.

This is from the Circuitmaker web site.

"Fourier Analysis – The Fourier analysis shows the frequency spectrum of any waveform. Simply click on a point in the circuit with the Probe Tool to plot the frequency spectrum at that point. You can define the fundamental frequency as well as the number of harmonics."


This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.