• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Replacing a cathode resistor with a CCS

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
EC8010

EC8010

Ex-Moderator
Joined 2003
Re: ((Re x Ie/UT)+1)/hoe = r output

oldeurope said:
Hello SY,
is the formula in #34 "too complicated" to calculate or is it just the "three wise monkeys" ;) ?
Click to expand...

Gently, Darius. Your equation breaks into two terms, one of which is 1/hoe, which I would expect. The other is:

(RE x Ie/UT)/hoe

Now, I'm not so happy about that term because it implies that you can claim hfe = gm/hoe and we all know that 1/hoe is particularly changed by Early effect in PNP transistors, yet a good CCS can be made with PNP transistors.

A much more usable equation is:

rout = hfe x RE + 1/hoe

Which is not so very different from your equation, but it uses hfe, which is at least easily measurable (often provided as a function on DVMs and semiconductor testers).
 
B

Burnedfingers

Member
Joined 2002
D

DrewP

Member
Joined 2002
Can any of those regularly frequenting CCS land offer an insight as to the relative sonic benefits of adding a CCS in place of a resistor in contrast to adding a B- supply and a bigger cathode resistor instead? (I'm thinking in an application such as a LTP phase splitter).

Cheers,

Drew
 
B

Burnedfingers

Member
Joined 2002
quote:
Originally #44 posted by burnedfingers
Now that we are done with the formulas ...


Understanding who is telling you porkies in the forum. ;D
quote:
Originally #44 posted by burnedfingers
...

Does anyone offer a ready made board for the CCS?

Take a resistor and donate your money for poor people, church, etc. .

D.

oldeurope,

Your sarcasism really isn't called for here. Granted I cannot keep up with some of your formulas and I'm not trying to. You need to take this thread as it was intended. You got to keep it simple for us uneducated people like me you know. Hell, all I wanted to know is would I achieve better results. But hey lets waste 2 pages to challenge a formula or better yet lets show the good people here that your just one step above them.



staggerlee,


I don't know what your problem is..... to drag up a post from another thread that doesn't apply here. The "Bottlehead" kit was brought up in the thread when we were discussing the possibility of a group buy for CCS boards. To be honest here if I wanted the damn thing I guess I would have bought it by now.

Geek,

Thanks for the reply. I'm just looking for a CCS blank board right now. I don't need the tube pads and don't have room for that large of a board.

Better yet I will forget this damn idea all together. Maybe its time for me to pull out and not post here again.
 
EC8010

EC8010

Ex-Moderator
Joined 2003
burnedfingers said:
oldeurope,

Your sarcasm really isn't called for here.
Click to expand...

I suspect you'll find (see post #43) that oldeurope was directing the irony towards himself for getting carried away earlier.

To attempt an answer to your unspoken question, "Are transistor CCSs really worth the bother?" I'd say yes, absolutely. In fact, I was surprised to find some years ago that I could even hear differences between CCSs (or at least, what should have been an adequate CCS and a rather better one). Myself, I like the cascode CCS, and I don't mind whether it's made of bipolar or FETs, but I do like to see a low output capacitance. I especially like to use them in your application of LTP phase splitters. A board is convenient, but a basic cascode CCS is so simple that it can be hardwired with transistors hanging in the air. Not very neat, perhaps, but technically superior to a PCB (less leakage and reduce capacitance).
 
SY

SY

Ex-Moderator
Joined 2002
SYclotron.com
EC, thanks for verifying the formula for output resistance.

The capacitance question is interesting. Let's consider a PNP cascode CCS in the plate circuit. The collector of the cascode device is connected to the plate, its base is DC-level-shifted from the base of the CCS device, which itself is DC-level-shifted from the supply rail, and both bases are at AC ground.

So it seems, on my naive inspection, that there is still a pretty significant capacitance there, and one that's modulated. On the plus side, the collector-to-base voltage is likely to be large, which should tend to minimize the modulation. But that collector-base capacitor is still there, or so it seems to me. Is there a flaw in my thinking?
 
zigzagflux

zigzagflux

Member
Joined 2006
burnedfingers said:
Better yet I will forget this damn idea all together. Maybe its time for me to pull out and not post here again.
Click to expand...


No, no: a CCS is well worth it, in my limited experience with LTP's.

I have some CCS boards that are adaptable to either a single device or cascode. Think I have two left. They are small (2" x 1.75"), and are designed for either a DN2540 or 10M45S device, in addition to being able to heatsink the upper FET for at least 5W, probably more if you don't mind running them hot (I build for 5 second "touchability"). Interested?

In my very short life as a toob nut, I have found these devices like as much voltage headroom as you can give them. My assumption was the lower capacitance issue SY is referring to. With lower headroom (like 5-15V) I have noticed significant sonic degradation.
 
EC8010

EC8010

Ex-Moderator
Joined 2003
Output capacitance

Yes, there's output capacitance between collector and base (bipolar transistor), drain and gate (FET), or anode and g3 (pentode). Plus stray capacitance to ground. I suppose if you were really clever, you could bootstrap that capacitance, but whether it would be worth it in the face of the strays to ground...

The FETs are the most problematic. Cdg is horribly dependent on voltage. There's nothing for it but to make the audio change in voltage as small a proportion of the DC voltage as possible - hence zigzagflux's comment about having as much voltage as possible. Worse, some FETs have some very odd effects at low voltages.
 
EC8010

EC8010

Ex-Moderator
Joined 2003
"Boing! Time for bed, said Zebedee."

SY said:
That argues for the use of something like a DN2540 cascode as an anode load but not as a cathode load.
Click to expand...

Yes, but capacitance isn't the only issue. It's possible to use DN2540 as an anode load quite simply, whereas bipolars tend to be more awkward. I'd probably be happier using DN2540 for an anode load and a nice fragile NPN bipolar for a LTP CCS.
 
zigzagflux

zigzagflux

Member
Joined 2006
But if we have, in a DC coupled SE gainstage to LTP, at least a hundred volts available at the tail, would the DN2540 cascode not be a great choice in the cathode? I don't get the improvement over using it as an anode load.

According to the datasheet, DN2540 reaches a low and linear 15 pF after about 20V Vds.
 
SY

SY

Ex-Moderator
Joined 2002
SYclotron.com
I assume you mean a self-biased cascode? In that case, a cathode loading means connection to the drain of the top device. The gate of that device is tied to the source of the lower device which will be very close to AC ground. So you have the reverse capacitance in series with the stoppers and a negligible term representing the looking-in impedance at the lower transistor's source.

A DN2540 is a particularly felicitous choice for a MOSFET here because that reverse capacitance is only a pF or two. That will probably be swamped by strays. Other devices have more like 10-20pF and maybe, just maybe, that could have a discernible effect. All of that is moot for most circuits where the impedance at the cathode is low, but for diff amps with high plate loads (like CCS), this could be significant.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom