So then what we have here, is a very, very bad CCS.
I call it a current mirror because that's what it does. It has a gain of three at the transistor's collector. Of course it'll act like a CCS if it has constant-current input. But because of global NFB, the currents in the entire LTP will be relatively constant anyways depending on the extent of the feedback.
So then we call it a CCS simply because it uses a transistor and a diode?
- keantoken
I call it a current mirror because that's what it does. It has a gain of three at the transistor's collector. Of course it'll act like a CCS if it has constant-current input. But because of global NFB, the currents in the entire LTP will be relatively constant anyways depending on the extent of the feedback.
So then we call it a CCS simply because it uses a transistor and a diode?
- keantoken
keantoken said:
sawreyrw said:
here is schema:
http://www.diyaudio.com/forums/attachment.php?s=&postid=1694415&stamp=1230110937
Thiis is a mirror, with 470 and 150 Ohm resistors. (not e.g. 100-100 Ohm)
Most mirrors have I_out = I_in
This one has like 3 times higher current output, which is a mystery????
I_out = (470/150) x I_in
One of the reasons to use current mirror in LTP pair is equal sharing of LTP current.
Both the input transistors should ideally run at equal current.
Thank you Lineup, Homemodder. I didn't know what to think.
IMO I don't think it's any use calling something anything other than what it does best when it comes to electronics. Otherwise, confusions like this occur.
Maybe this causes dominant 2nd order harmonics? Perhaps it was intentional.
thanks,
- keantoken
IMO I don't think it's any use calling something anything other than what it does best when it comes to electronics. Otherwise, confusions like this occur.
lineup said:
Maybe this causes dominant 2nd order harmonics? Perhaps it was intentional.
thanks,
- keantoken
As andrew T said, there is no current mirror, I gave daniel
the 2 simplest versions of my amp to get him started ( blameless
without CM and a bootstrap version).
But at only .$18c for 2 bc556's, he can feel free to add it.
and the cost difference of the lower power version 30-0-30ac
VS the 45-0-45ac power supply is at most $20 (caps and trafo),
so for a minor expense you can have a full blown 4 OP device
125/200W "blameless".. Unless ,as I first assumed, you had
power supplies left over from your IC endeavors.
OS
The standard two-transistor current mirror relies on the parts being similar in order to provide mirrored output.
This one relies on the matching of the transistor and the diode to provide mirrored or proportional output.
There is actually no or little difference between the two circuits.
The fact is that it DOES mirror current, and so I see nothing wrong about calling it a current mirror.
The problem with a discussion (or debate) like this is that both contenders try and "educate" the other based off of the conclusion that they are right. Similarly, the debates usually get nowhere. I don't want to continue this.
- keantoken
This one relies on the matching of the transistor and the diode to provide mirrored or proportional output.
There is actually no or little difference between the two circuits.
The fact is that it DOES mirror current, and so I see nothing wrong about calling it a current mirror.
The problem with a discussion (or debate) like this is that both contenders try and "educate" the other based off of the conclusion that they are right. Similarly, the debates usually get nowhere. I don't want to continue this.
- keantoken
keantoken said:
I call it one current mirror, too.
Even if there is a 'fixed current gain' output.
Often such current mirrors are used in no-global-feedback amplifiers.
To increase the current level in next stage, without using 'normal gain transistor'.
The output will be a fixed number higher current.
The constant, fixed relation of current is set by selecting the 2 emitter resistors.
I guess you could call the CM a bad CCS but it augments
the MAIN CCS which is very good . It (the led) runs at 5ma,
which by itself has .1ma deviation at 2V ripple on rail. Divide that
by the gain of the CCS tranny and you end up with a rock
solid .001ma deviation at the LTP Re's..
OS
Attachments
Alright, problem solved. I agree, MJL. But in my experience if someone is a discerning person looking for answers they will usually find the answer when they can think easier, after the debate has ended. Because of this I opt for the most peaceful solution. It doesn't matter if someone realizes their mistakes immediately. But nothing in life comes immediately when we want it, so we must have patience with our fellow neighbor. I am glad we were able to end this peacefully.
Uh-oh, we must look out for our friend AndrewT! Look, this is not an isolated phenomenon!
http://www.diyaudio.com/forums/showthread.php?postid=1691856#post1691856
thanks,
- keantoken
Uh-oh, we must look out for our friend AndrewT! Look, this is not an isolated phenomenon!
http://www.diyaudio.com/forums/showthread.php?postid=1691856#post1691856
thanks,
- keantoken
keantoken, I could be wrong but I dont find it to affect the harmonic spectrum. I have a design which I have used for 14 odd years now which uses this in the vas and I just had a quick look at how it sims. Better THD figures using this circuit instead of mirror, by a small margin but there. Both 2nd and 3rd down. I cannot remember exactly why I chose it instead of mirror besides sounding better at high volumes. Im looking at a module now and I see that I also trim the resistor value to get exact balance. Back when I designed it it was with scope and ears.
Most transistors, except for the 2n5771/5769, have a Cjo of around 10p-20p. A 1n4148, however, has just 4pF of capacitance. This may have something to do with this. At high frequencies this would be relevant, if it were in a part of the circuit with high open-loop gain. If the Cjc is too much greater than the Cje, it can actually cause problems at HF, and so a simple 100 ohm resistor might be put at the collector of such transistors. This goes for most transistors in a circuit that have a role in gain or OLG.
Also, a transistor has an extra p-n junction. A diode does not.
I don't know how much of an impact this has on distortion, maybe I will test in sim.
But I think that diode are more rugged than transistors, less complex, so perhaps this has something to do with it.
Thanks for the info, Homemodder.
- keantoken
Also, a transistor has an extra p-n junction. A diode does not.
I don't know how much of an impact this has on distortion, maybe I will test in sim.
But I think that diode are more rugged than transistors, less complex, so perhaps this has something to do with it.
Thanks for the info, Homemodder.
- keantoken
The impact is very small, but Im going to have a second look at its clipping behaviour in the vas as I suspect this would have something to do with the sound at high power. Just wondering why Ive never run into clipping related problems with this amp as it does 500w and the modules are for car subwoofer use.
BTW if you interested a fellow member here tried out that little current source in one of his designs, it had quite an impact, more than I usually find, I could email you his results.
BTW if you interested a fellow member here tried out that little current source in one of his designs, it had quite an impact, more than I usually find, I could email you his results.
keantoken said:
Attached here (click photo once or twice) is a short brief on this current mirror. I don't understand it too clearly, but the lights are beginning to shine through the fog. This brief is from the same document as post #13.
For me, how it works is a bit less important than the explanation of its role (which was already provided).
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.