| Magura |
The initial question in the old thread was never really answered....it turned into source follower and the like, so id like to ask the initial question once again. Maybe the answer is obvious now after the development in the other thread....
It got as far as we agreed that 50mH chokes were a good idea for replacement of the bias resistors in the SOZ.
Now for the less talented electronic constructors there are a few questions that were left open.
Would it be clever to get as low resistance in the choke as possible? Or would it come to a limit where other factors make a difference?
I saw a question that were left un-answered on the old thread, if its possible to replace all the 8ohm resistors with chokes? (i guess id just lower the values for the other resistorsbut.....).
Magura |
|
|
| joensd |
Sorry to get offtopic again but it´s not possible to email you.
You were talking about doing some chokes yourself.
Do you have any offers already?
Might be interesting for others as well for a small group buy.
Please let me know if you do.
TIA
Cheers
Jens |
|
|
| Magura |
Please try email again.
The email i use for the forum is :
travelling_mac@yahoo.com
It should work through the forum though.
Yes, it would be no problem to wind a few inductors for the members of the forum.
Magura |
|
|
| Magura |
Havnt anyone got any actual experience with choke loads for SOZ at this point???
Sounds strange, since the choke load thread is fairly old.
Cheers
Magura |
|
|
| Kashmire |
When my Zen v.4 boards arrive from Pass DIY, I plan to listen to them "stock" for about a week. Since I have a regulated power supply, I will remove Z1-Z5 and R19 and ground C11.
Afterwards, I will replace Q2 with a choke and reduce the power supply voltage to 25V (to maintain 20V across Q1). I'm assuming that 4 volts will drop across Q5 and 1 volt will drop across the choke. R4 will adjusted to maintain 2A bias current. I need to study the circuit to be sure Q1's bias will function properly in this configuration, but I' also like to remove R0 and R1.
I will post my findings.
The choke will be a toroidal power choke rated at 3A static DC. I will order 10 of these at $50-$55 each. I will be interested in selling some of these.
As mentioned in the other thread, I will also be comparing the Zen v.4 with a source-follower amplifier:
http://www.diyaudio.com/forums/show...&threadid=24737 |
|
|
| Nelson Pass |
Choke loads for SOZ are great. Use some value of resistor,
like 22 ohms (at the appropriate wattage) across the choke
to help keep it damped and stable. |
|
|
| Magura |
ok, so choke loads it is. Wich resistors can be replaced with chokes? and can i just make the rest of the 8 ohms resistors the same value as the resistance of the chokes??
What about the 1 ohm resistor....it should not be changed, right??
Cheers.
Magura |
|
|
| eLarson |
I'm thinking you'll just want to replace the 8-ohm power resistors that serve as the drain loads.
eL |
|
|
| Magura |
But dont i have to reduce the voltage as i reduce the resistance by replacing the resistors with chokes??
If so, wouldnt it be nessecery to reduce the values of the other resistors to maintain the same current?
Cheers
Magura |
|
|
| GRollins |
I'm not sure which 'old thread' you're talking about, but I don't think you'll like the results if you use inductors under the Sources of the MOSFETs in a SOZ.
Inductors as loads, on the other hand, will work just fine.
Inductors should have as low a DC resistance as possible. In the real world you won't get to 0, of course, but you should try. In a high current circuit such as the SOZ, the coils' DC resistance will make a difference in how cool the coils run. Use air core inductors. Ferrite or iron core inductors (regardless of construction) will tend to saturate since the DC current is already generating a large magnetic field. When the AC current--aka music--peaks, the long-suffering core will not be able to accept any more flux and the signal will be distorted. Tube folks deal with transformers because they have to, not because they want to. There's no reason, barring masochism, for solid state people to inflict such things upon themselves. Kinda makes you wonder about McIntosh and their autoformers, but that's another topic for another day.
No, you wouldn't replace all the resistors in the SOZ with inductors, just the load resistors (as eLarson notes)--the ones above the Drains of the MOSFETs.
(It seems that there's an echo effect around here. When a question comes up in one thread, you frequently see the same question in another thread. Almost exactly the same questions came up in the SOZ w/current source thread a week or two ago. For what it's worth, the answers haven't changed. Oh, well.)
The one ohm resistor stays.
The resistor values have no relation to the DC resistance of the inductors.
The 8 ohm resistors under the Sources of the MOSFETs determine the bias current. With the Gates at ground potential, the Sources will settle in at about -4 volts, plus or minus a bit. The Drains will then be connected to the inductors, which are connected to the positive rail. The larger the value of the inductor, the more output you'll get, and the more bass. A really big inductor will give you the same amount of bass that a resistor would, but with less heat for the same power output. How big an inductor? Infinite. Ugh. That gets difficult to deal with. Your fallback position is simply to buy or make the biggest inductor you can practically deal with and go from there.
Grey |
|
|
| Magura |
| quote: | Originally posted by GRollins
The resistor values have no relation to the DC resistance of the inductors.
The 8 ohm resistors under the Sources of the MOSFETs determine the bias current. With the Gates at ground potential, the Sources will settle in at about -4 volts, plus or minus a bit. The Drains will then be connected to the inductors, which are connected to the positive rail. The larger the value of the inductor, the more output you'll get, and the more bass. A really big inductor will give you the same amount of bass that a resistor would, but with less heat for the same power output. How big an inductor? Infinite. Ugh. That gets difficult to deal with. Your fallback position is simply to buy or make the biggest inductor you can practically deal with and go from there.
Grey |
Thanks Grey, this was just the kind of answer i needed. I need this spelled out, im new to audio electronics of this level.
So if i have undrstood this right, i get 4 times as much output if i use an inductor of 2 ohms resistance??
Isnt there a way to calculate the frequenzy response VS. inductor value?
Cheers.
Magura |
|
|
| Magura |
Could it be possible to draw a paralel from passive crossovers??
I mean the formula for calculation of inductor values??
In order to make a tri amp version, with only the need to limit the upper end frequenzies....instead of a regular X-over.
Cheers
Magura |
|
|
| woody |
Refering back to the SOZ circuite the 8 ohm drain resistors
determined the output resistance and damping factor. Now
if we put a pair of inducters to replace the restors how
would we figure the outout impedince ? |
|
|
| Magura |
Why would the impedance change??
Cheers
Magura |
|
|
| azira |
If you replace the source 8-ohm resistors with a choke, you'll be basically shorting out your power supply... prepare for a lot of current to flow.
If you have 2-ohms DC resistance from the choke, you'll have effectively a 1-ohm resistor there and a total of 3-ohms from source to -Vee on each FET. That will bias the FETS at 3x the current they were biased at.
Replacing the drain resistors with chokes will give a high impedance for AC which means higher gain and efficiency. I think that you would want some resistance behind it though to set the DC operating point.
Putting a choke in series with the source 8-ohmers should improve CMRR, it will not affect output impedance or gain.
Just my thoughts about the diff-pair analysis...
--
Danny |
|
|
| Tyimo |
| quote: | | (It seems that there's an echo effect around here. When a question comes up in one thread, you frequently see the same question in another thread. Almost exactly the same questions came up in the SOZ w/current source thread a week or two ago. For what it's worth, the answers haven't changed. Oh, well.) |
Hi!
I was who asked about to replace the resistors! :-))
I would like to build the SOZ as Grey suggested: "replace just the load resistors, the ones above the Drains of the MOSFETs."
-And now I would like ask how can I calculate the PSU achievement for a 15 W SOZ?
-What is the power rating of the rest of the power resistors? (50 W?)
-Would the small resistors value the same as in the original article? (R8-11)
Thanks! |
|
|
| Magura |
By doing the choke thing, it would be possible to exclude the capacitors from a passive X-over?? For tri-amping that is !! only inductors in the X-over. All thats needed is to determine the cutoff frequenzy of the amp VS. inductor value. I guess it will have to end up as an experiment :(
Cheers
Magura |
|
|
| azira |
hypothetically.....
If you had a setup where you highpass the tweeter, highpass the mid at a lower freq and fullpass the woofer by basically doing a choke to gnd in parallel with the speaker... then you could "technically" rig your chokes to the right XO freq and not need them in the speaker.. i think...
I bet it would be pretty tricky though. You'll need to determine your output resistance of your driver pretty tightly. probably... if we're talking an SOZ. |
|
|
| Magura |
What worries me more is that none of these brainy people around here got an answer to how to calculate the relation between inductor value and freq. response......i really dont like that, since it smells of loads of experiments......winding loads of huge inductors.
Cheers
Magura |
|
|
| azira |
Sorry... first order filters was basic analog circuits... here's the equation.
Corner freq for a LR is = L / ( 2 * pi * R ) ...
So... from the output node you need to know the resistance to gnd (that's signal ground not DC gnd) and inductance to gnd to figure this out. Since your R is fixed from your load and bias conditions (FET output resistance), all you need is to calc is L...
But since your R is relatively small, and in the SOZ case, it's 1/2 R... you can sort of ignore the FET output resistance. So your R is just the speaker... |
|
|
| Magura |
What would typical R be of a 10 inch 8 ohm driver???
Cheers
Magura |
|
|
| azira |
| You just answered your own question :) R is resistance as in Ohms... So 8-ohms. |
|
|
| Magura |
yes....but dosnt it change with freq. ??
Cheers
Magura |
|
|
| azira |
Well.. yeah, but that fundamental problem hasn't really changed since this thread started, it's sort of an inherent idiosyncracy involved with passive XOs to begin with...
The 8-ohms is basically the average impedance in the flat region of the speakers response. The speakers impedance rises sharply at lower frequencies until the resonance freq and then falls off again.
So... just use the 8-ohms and you should be ok (meaning 'close enough'). If changing speaker impedances worries you, look into zobel networks to counter it..... and then you'll be left with a real 8-ohm resistance that you'll do the exact same calculations to and get the exact same result. |
|
|
| Magura |
how do you fill in the formula then to get the freq. response/ inductance relation ??
Cheers
Magura |
|
|
| GRollins |
Actually, the resistance for an 8 ohm driver is typically on the order of 6 or 7 ohms, the remaining 1 to 2 ohms being comprised of the inductive component. Or to put it another way, if you use a regular meter to read the 'resistance' of a driver, you'll find that it's less than the nominal impedance, since meters use DC to measure resistance.
Speaking for myself, I've always kinda wished that they hadn't used the term Ohm for reactive components, but it's too late to change it now.
Grey |
|
|
| Fuling |
Is it really such a good idea to use the chokes as crossover components?? When the frequency goes down the loadline for the Mosfet becomes pretty much vertical, which I guess has negative effects on the distortion.
Thinking of it, this would be equal to an amp driving a crossover that short circuits the lower frequencys instead of blocking them.
I would have used active crossovers (possibly integrated in the front end of the amps by adjusting the input caps) and chokes that could handle at least twice the estimated frequency range of the filter section. They would still be quite small for the mid/high sections. |
|
|
| azira |
Interesting perspective Fuling... I suppose I wouldn't necessarily recommend tuning the amp chokes myself but more because you would want an amp to be as general purpose as possible and then tune the loading.
However, consider this, the choke will actually have something like 2-ohms of dc resistance that is passing the bias current. This resistance can only go higher with frequency which is a good thing. So when we get low enough where the choke starts to look like a short circuit, it'll really just be passing the bias current like it already has to do. |
|
|
| Magura |
There is a point to this.
The inductors for the upper 3rd of the spectre would be just fine with a few mH, like 15-20 mH (assuming that mid would be something like 50mH and bottom would be like 80-100mH).
In that case the resistance of the inductor is easily below 1 ohm !!!
Maybe its smart to not make any inductor less than 2 ohm.
Cheers
Magura |
|
|
| azira |
I hate how I keep thinking about more stuff everytime I read a response that makes me feel like I should have put more thought into my previous posts....
Here goes....
The bias current actually comes from the source resistors and not the drain anything. The main reason for the drain resistors is to generate voltage swing since there will be current swing. But even if the drain resistors were replaced with a short circuit, you would still certainly get a varying current from the supply rail through the transistor.
So... with a choke, we're just adding impedance for AC but allowing DC to pass through it... it's actually better than using a coupling cap because your choke will produce a much higher parallel impedance at AC with a low impedance DC path.
I think the choke can have a low DC R and still be just fine. And to that end, yes you can vary the rating like you suggest.
This brings up another potentially interesting topology maybe better saved for another thread... what about putting the load in parallel with the choke in a single ended application... |
|
|
| Kashmire |
Indeed, the load can be placed across the choke instead of the active device. If you draw the circuit, you'll find that this topology puts huge demands on the power supply. In addition to supplying the steady-state bias current, the PSU must also supply transients. This means the PSU acts like a Class-AB power supply, requiring lots of energy storage and you can't use a "capacitor multiplier" circuit.
In this topology, the PSU capacitors become part of the signal path.
Putting the load across the choke also routes any PSU noise into the load. The choke will resist the noise, thereby forcing the noise into the load.
These topics are addressed in an existing thread regarding choke-loaded amplifiers:
http://www.diyaudio.com/forums/show...8719#post288719 |
|
|
| Tyimo |
Hi!
Do we speak about something like this?
The inductors are centertaped toroid coils. The suply voltages are may be a bit low, but could give some output power near to 10W. (?) |
|
|
| Tyimo |
Is there anybody out there?!?....
Mr. Pass?Henrik??
This thread is dead, because nobody like the SOZ anymore..... :((( |
|
|
| Nelson Pass |
Not much to say. Your schematic looks fine to me.
I think that current sourcing on the order of Figure 8
in ZV7 will give you more efficiency and betther operation
with a single-ended input. Also, the 22 ohm resistors
across the coils are probably better if you send them to
ground, which will give you a little better PSRR, |
|
|
| Tyimo |
Mr Pass!
Thanks a lot for your answer!
I would like to stay by the original SOZ biasing methode (with power resistors) because I have heard that it give nicer sound than with CCS/ACS. Ofcourse it is a matter of taste..,but I would like to try it and I will use ballanced input from my SOSOZ anyway.
So, same more question for this circuit:
-what is the best place and Wattage-value for the 22 Ohm resistor? Should it stay as it is?
-What is the correct PSU Voltage for a 10W inductor loaded SOZ?
Thanks!!
Tyimo |
|
|
|
