X-lite

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Note: This IS NOT a proven design. If you build it, it will be at your own risk. All info based on simulation:
If you are familiar with the AlephX, you will see that there is nothing new here- just re-hashing some old ideas into something different. I was asked to write up a description and include a readable schematic (see the zenlight thread). So, if you're interested- here goes:
refering to the attached schematic (zip file)
The input stage:
Q1 and Q2 form a differential pair, biased by current source mosfet Q3. R13 (1.21K) is there with C1 to divide the rail and decrease dissipation on the current source mosfet. D1 sets up a 9.1V reference, filtered by C2, and R22 sets the current. R22 will also determine what voltage the speaker outputs of the amp are at, and should be set so that both outputs are at +3-4 volts for maximum voltage swing. R12 sets the current through the zener D1 to around 4mA, and should be adjusted if the + rail voltage is changed. R29 and R2 are the load resistors for the differential pair. R25 (2k pot) is used to adjust the balance between the output stages, so you can null out DC voltage across the speaker terminals caused by imbalance between the output stages. For those who are experienced with the AlephX design, you may also substitute the HH current source.

The output stage:
L1-L6 are 110V 250W incandescent light bulbs, as found everywhere in the US. They are used with the high + rail to create a current source. I used Nelson's chart in the Zenlight article to approximate the resistance of each bulb for the simulation. Your results may vary... (you have been forewarned- this is not a proven design at this point). The output mosfets can be whatever you would use in the zen/soz/aleph/alephX amps... typically IRFP240. The .3 ohms resistors are there to help balance the mosfets, so that any one dosen't hog all the current. There are the typical 221 ohm gate resistors to prevent oscillation, as used in all the aleph amps, etc.

Other stuff:
R14 and R15 are the "magic" resistors, which you know about if you have been following the AlephX thread. They stabilize the amp. Along with these, you might want to use 33-100 ohms at each speaker output terminal to ground to help with stability.

Some things to note about the design:
The power supply will be massive- needs to supply 10+ amps at each rail, all the time. It should do about 45 watts into 8 ohms, and should drive down to 6 ohms (actual impedance- not rated speaker impedance) without current limiting. There is about 150 watts dissipation among the mosfets, so a good heat sink is required.
If you don't need as much power, you could get rid of one bulb and one mosfet on each side. The reason I decided to go with so much bias is to guarantee that it won't current limit before reaching max voltage output into 8 ohms. If you don't need a lot of power, you could get away with less bias.

Any suggestions or improvements are definitley welcome!

Steve
 

Attachments

  • x-lite_schematic.zip
    36.2 KB · Views: 648
diyAudio Editor
Joined 2001
Paid Member
Wow! Now that the schematic is clear, I'm even more impressed. Thanks for adding the resistor ratings, etc.

This one has it all! Like a walk through the history of Pass DIY
:Pirate:

x-circuit :) ,lightbulbs :bulb::bulb::bulb::bulb::bulb::bulb:,
decent power output :p, huge power supplies :grumpy:, simple circuit :spin:

How do you get the 12v on the negative supply vs. 50 on the positive?

Thank you for sharing, maybe the mods will move some of the other postings regarding this amp over to this thread.
 
How do you get the 12v on the negative supply vs. 50 on the positive?

The 12V on the negative supply is to keep the dissipation in the output fets down. The 50V is needed to make the bulbs act like current sources- basically a somewhat high voltage and a resistor, like in the balanced line stage.
These rails will have to come from seperate transformers, with the + of the 12V rail tied to the - of the 50V rail to make a ground reference point...
So, 2 transformers, 2 rectifiers, 2 sets of caps, etc.
Thanks for the interest.
Steve
 
diyAudio Editor
Joined 2001
Paid Member
Thanks Steve . I was afraid that might be one of those stupid questions! I guess it was OK after all. I know I should do SOME of the work, but any clue on the VA of the big transformer and the 12 volt one? I assume they are different loads.

Hey guys, go to e-bay and search on "projector bulbs"
Some VERY cool options- including 750 watt 120 volt models
I really hesitated mentioning this, cause I'm not ordering yet!
 
mrothacher,
I don't think that split supplies are a necessity for X feedback, but i think the circuitry used generally lends itself to that type of setup. Nelson could tell you for sure! How's that for answering but not saying anything! Sorry I can't get more specific.

Variac,
No stupid questions, just stupid people, right? haha Just kidding. Actually, my sig is directed at myself more than anyone else. I have been known on occasion to ask a question that 1. I already know the answer to, but have not taken time to think through clearly, or 2. That is just plain stupid in general. I should listen to Mark Twain more often.
The big tranny will probably have to be 40 to 45 volts @ a bunch of amps. Think somewhere around 750-1000VA minimum! Per channel! Yikes!
As for the 12 volt- a filament tranny with a lot of current should do nicely. I know allelectronics has one for sale (they call it 12.5 volt @ 25A). I'm using some of these for my Aleph X's. The currents will be the same for both trannies (or, should I say the draw on each rail will be the same). The upper rail will have to be a lot higher VA because it is supplying the same current and a lot more voltage.
Are you actually thinking of building one of these monsters?
Steve
 
diyAudio Editor
Joined 2001
Paid Member
I am definately in the lazy question asker camp ;)
Didn't Stefano say he has actually built one of these?

Well... I have these 2 1000 VA toroids @ 37.5 volts that I was planning to use for an Aleph 2. BUT the x amp designs came out, so I wanted to be trendy, but they use about half the voltage so no go. But your amp seems like it would be OK with the 37.5 volts. Then the problem of all those expensive heat sinks which your design at least partially addresses. The biggest problem is I have a 15 watt SOZ in the works for the treble so the combo of the 2 amps would REALLY suck up the power-maybe more than my house wiring can handle!
 
maybe more than my house wiring can handle

You are not kidding! I can see how you could easily overload the circuit with a biamp setup like that. I hope to find out soon. I wonder when Nelson's new amp will be out at Passdiy? I've been watching non-stop for it.
Well, if you decide to give it a go and I can be of any help, I'll be around. You and I may be the only ones to build one, from the looks of it. Oh well... the rest might just miss out on something great!
Steve
 
Variac said:

Didn't Stefano say he has actually built one of these?



Yes but single ended .I am drawing my own schematic for the X-lite yet.
But I will use much less power:dead: than a kilowatt/channel ...
let's say 4-5 Amp par channel max ... :cool:
And 2 mosfets par channel .Also becouse 4 heatsinks are just waiting for it.
And I've got 2 tranny with 37-0-37volt and 15-0-15 each.
Actually they are connected for 57volts.


Thanks guys!
 
stefanobilliani said:



Yes but single ended .I am drawing my own schematic for the X-lite yet.
But I will use much less power:dead: than a kilowatt/channel ...
let's say 4-5 Amp par channel max ... :cool:
And 2 mosfets par channel .Also becouse 4 heatsinks are just waiting for it.
And I've got 2 tranny with 37-0-37volt and 15-0-15 each.
Actually they are connected for 57volts.


Thanks guys!

Great! I can't wait to see how it turns out! Are you sure that you want to lower the bias so much? I have found that it is already fairly current limited with the monster tranformer. The Aleph X gets around this with the active current sources, but with this design you are in need of A LOT of bias. I guess it will be alright if you're going for a lot less power. Make sure you post your schematic, if you feel like sharing it. I would like to see what you come up with.
Steve
 
Thanks Steve!

I just finished my work week and rushed to the computer to see what you have for us. Wow, that's a very nice schematic you've drawn up. I'm going to give it a thorough looking over.

I happen to have two 1000 V/A 52-0-52 volt Plitron trannies for the project. With an L/C filter I should get pretty close to the 50 volt requirement for the positive rail.

I can't thank you enough for you efforts. Here's to great listening sessions!

John Inlow
 
nobody special said:


Great! I guess it will be alright if you're going for a lot less power. Steve



The single ended version doesn’t seem suffer from current limits playing music.
I also put at the input some sordid sinusoid signals at 16 and 32Hz , the amp manage they without problems . The speakers are 94db/1W/1meter but also 3 ohm of impedance at regions like 200-300Hz.
The contest is encouraging . For the balanced version I tend to think at the double of the power consumption… just for the start!


Oh… it is time to go to buy some more halogen sticks ;)
 
stefanobilliani said:




The single ended version doesn?t seem suffer from current limits playing music.
I also put at the input some sordid sinusoid signals at 16 and 32Hz , the amp manage they without problems . The speakers are 94db/1W/1meter but also 3 ohm of impedance at regions like 200-300Hz.
The contest is encouraging . For the balanced version I tend to think at the double of the power consumption? just for the start!


Oh? it is time to go to buy some more halogen sticks ;)


You have to remember that when you are driving a speaker in balanced mode that each side will be seeing 1/2 the actual impedance.. so you end up needing 4X the current. Glad to see that you're building it... can't wait to see your results. No contest really.
Steve
 
Re: irf9510 vs. irf9610 transistors

carpenter said:
Hi,

I have a stash of irf9610 transistors and would like to use them instead of the irf6510 as listed on the schematic. The irf9510 is capable of handling twice the current than the irf9610, although the 9610 handles twice the voltage. Is the 9510 mandatory?


John

John,
The input differential pair is only seeing something like 11mA each, so power isn't really a big deal. You will want to look at the input capacitance, Ciss, for any mosfet you might want to use. In your case the 9610 actually has less, so that is a good thing. You want as little as possible to keep the high frequency -3db point as high as possible.
Yes, there should always be gate capacitance- I accidentally left off the 221 ohm resistors. Sorry 'bout that.
Steve
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.