 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
182.239.90.5
Just want to know the general comments on DN2540 CCS, most of the time the web shows cascade, but what about just using one DN2540, as good ?
What is your experience？
Follow Ups:
Cascode for sure. The dynamic impedance of the top device is multiplied by the current gain of the device on the bottom. Note: the device capacitance will start rolling this impedance off at ~500Hz at 3dB/octave. Another reason to cascode, stacking two devices in series will reduce series capacitance. Do consider using a different MOSFET from the DN2540. I would recommend the IXYS 8N100D2. They sound better to my ears. Also, try to arrange for at least 40V across the CCS. These CCSs will "work" in single digit mA ranges, they work better over 10mA and best over 20mA. For currents less than 10mA consider using a different CCS.
"It is better to remain silent and thought a fool, then speak and remove all doubt." A. Lincoln
JKT,
How do you think the following post regarding adding one more resistor in series to CCS, does it make sense?
I've used CCS for plate loads on small signal tubes. I never added a resistor in series with the CCS so can't comment on the result. If you do try it, it will eat into the available B+ (V=I*R). Choke anode loads are very useful when the available B+ is limited. That said: quality chokes aren't cheap. A Lundahl anode choke with an amorphous core will set you back a pretty penny. Two MOSFETs and a couple of resistors will cost less that $5. My best advice is to experiment and try a resistor load, then a CCS and then a choke load if you can afford it.
"It is better to remain silent and thought a fool, then speak and remove all doubt." A. Lincoln
From Gary Pimm. Cascode is so much better.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 04/19/24
cascade is something different.
And, whilst in there, use a DN3545N3 'under' either a 10M45 or a DN2540N5. The former is preferred for a couple reasons... :)
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
If I'm doing 10ma through a plate load CCS does the DN3545N3 under a 10M45 need a heat sink?Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 04/19/24 04/19/24
The 'under' element has just the upper element bias voltage across it drain-source. Sooo...with 3V for the DN3545N3 to hide under at 10 mA, dissipation is 30mW. IOW, I really doubt it will need a heatsink.
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
The 10m45s have this nice chart on the datasheet but I don't see that with the DN3545N3.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
The datasheet linked through the digikey site has two sets of plate curves. the more detailed set has D-S voltage from 0-10V. It is not the bias R sort, but just as you'd use a set of screen-grid-less pentode curves to determine bias R, you should use those...
These are MOSFET's, not dedicated current regulators, yes?
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Yes, MosFets
sorry for being dense but how does this help me?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
We are hunting for a current setting resistor, yes? Whilst the 3545 is running beneath a 10M45, right?
So...draw a vertical line at 3V.
Then join me in the guess that we want about 1.7V for operation at 20-30 mA.
Apply Ohm's Law to determine the resistance required to deliver 1.7V at this idle current. Unless you wish to run 50 mA...in which case, note there is a gate line at that current( the -1.5V lowest one ).
Since the 1.7 is a guess/interpolation be prepared to adjust this value a bit. I suggest some battery bias to bring the effective bias voltage up to ~12 so most of the resistance is dedicated to covering the battery's extra voltage.
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Yes and yes.
I get it but I'm doing 11 ma. for my drivers and 10ma for the second stage of my phono preamp and 2.3ma for the first stage of my phono preamp.
I just used trim pots to find what value Rset resistance I needed in each instance. It's all working fine now. Thanks.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
.
Have Fun and Enjoy the Music
"Still Working the Problem"
.
Have Fun and Enjoy the Music
"Still Working the Problem"
What is those capacitance 32.4P, 0.177P ?
Thx in advanced
Hi, I was asking those small capacitance denoting what properties
Input capacitance. Think of it like Miller capacitance in a tube. That capacitance has to be driven. If the drive impedance is too high then the high frequencies will be lost.The lower the capacitance is the better the part is for making a CCS.
In cascode the capacitance goes way down. That is why cascode is better than single.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 04/11/24
On spec, active SS CCS provide 100M ohm load resistance that no real choke could exist that value
But in term of human listening, choke or CCS?
"...in term of human listening, choke or CCS?"
That is subjective. You have to answer that for yourself. No one can answer that for you.
Back to the objective facts. A triode loaded with a CCS will have more gain and less harmonic distortion than the same tube loaded with a plate resistor. Period.
A choke has reactance so at the frequency extremes it will not offer as high a load impedance for the tube like it does in the mid-range.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"CCS provide 100M ohm load resistance that no real choke could"
Resistance measurements don't communicate the non-linear idiosyncrasies of the SS junction within the device. In addition, a choke can significantly increase dynamic range (headroom) due to energy storage. No CCS can do that.
"non-linear idiosyncrasies of the SS junction within the device"
The CCS is such a high impedance vs. the plate resistance of the triode that it is loading that the non-linearity of the device will have very little influence over the total. A tiny fraction.
If the non-linearity of the MosFet causes a 10% decrease in the load as seen by the triode then you have say 9 megs vs. 10 megs. The distortion and gain of a triode loaded with 9 megs vs. 10 megs is right next to no difference at all.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"The CCS is such a high impedance vs. the plate resistance of the triode that it is loading that the non-linearity of the device will have very little influence over the total. A tiny fraction."
You can't know that. Impedance is a RMS measurement. It does not communicate the extent or duration of any non-linearities that may be created by the device under test.
But regardless of the duration of the non-linearity, that non-linearity would have to substantially lower the impedance of the CCS or it won't make an audible difference. Or at least I don't see how it would.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"But regardless of the duration of the non-linearity, that non-linearity would have to substantially lower the impedance of the CCS or it won't make an audible difference."
No one knows how much (or which type) of SS non-linearity is responsible for degrading sound quality. Any SS device that isn't bypassed introduces this possibility, and I simply won't take the risk.
There is magic in that box.
What kind of magic?
Don't know, but it may be bad.
............................................
Just trying to round up the 'omg it is solid state and that can be bad' comments... :)
First time I built a LTP is was a long resistor. Then tried a current reg diode. Then a cascode CCS from Gary Pimm. Then thought the simple single depletion mode would be darned near as good and *FAR simpler. Nope. Cascode depletion mode came next and that worked well, just like Gary's enhancement-mode stuff worked.
The 10M45 has 2x the bias voltage vs a DN2540N5. A very useful thing to provide the DN3545N3 I prefer in the bottom spot.
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Spec them right and they will never fail, not pick up any stray magnetic noise, not distort on overdriven peaks and are super simple to implement. Tubes were designed to be loaded by resistors.
If you need more gain, use a tube with more gain. ;-)
If I wanted semiconductors attached to the signal path, I wouldn't be working with tubes. That aside, this is excellent info to have when a CCS is needed for other purposes, such as test equipment.
I second TK.
Give me a choke. Or a resistor. :)
They are not really in the signal path. They are just making up a high impedance load. Any nasty SS distortion will only cause a small disturbance in the impedance of the load for the triode. That impedance, even at it's lowest point, as caused by the nasty SS distortion, is still much higher than any resistor would afford and still high enough that you won't know the difference between the highest impedance of the CCS (when it's not distorting) vs. the lowest impedance of the CCS (when it is distorting).
So, even though it is the plate load, those distortions are not really having any effect on the output signal because, SS distortion or not, the impedance of the CCS is so much higher than the plate resistance. Therefore it can be said that the CCS is not "in the signal path".
Years ago Mark Kelly explained this much better than I. He spoke in terms of the impedance being so high that it had very little "grip on" the circuit.
I think it's just a matter of looking the the impedances involved and doing the math.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
