This design will be very sensitive with temperature, and much worse, OLG will be highly temperature dependent.
Nelson, try to make first a simple CCS as per your design and watch what is going with output current when temperature will change when your finger touch the jfet
This is a folded cascode design ... the LTP current MUST be very stable.
Nelson, try to make first a simple CCS as per your design and watch what is going with output current when temperature will change when your finger touch the jfet
This is a folded cascode design ... the LTP current MUST be very stable.
Hey guys.
I have a question regarding this circuit.
How do you calculate how much current is running on the folded cascode VAS stage?
Let's say we have 4mA of long tail current pair through the input stage. And then we have 160ohm of source resistors on the JFET sources. So we'll theoretically have the 4mA split into two into 2mA and 2mA each flowing through the resistors.
Since there's a split after the resistor, one to the collector of the cascode and another to the input stage, how does the bias diodes of the folded cascode stage interact with the long tail pair CCS and the source resistors so we can tune the total amount of current flowing across the vas stage?
Thanks.
I have a question regarding this circuit.
How do you calculate how much current is running on the folded cascode VAS stage?
Let's say we have 4mA of long tail current pair through the input stage. And then we have 160ohm of source resistors on the JFET sources. So we'll theoretically have the 4mA split into two into 2mA and 2mA each flowing through the resistors.
Since there's a split after the resistor, one to the collector of the cascode and another to the input stage, how does the bias diodes of the folded cascode stage interact with the long tail pair CCS and the source resistors so we can tune the total amount of current flowing across the vas stage?
Thanks.
TzeYang said:
Please read this:
http://www.diyaudio.com/forums/showthread.php?postid=119592#post119592
Jonathan CARR is a folded cascode guru.
I got the pot, very handy, thank you very much. I also found an AD797 model and other candy there as well. Great!aparatusonitus said:
I'm sorry to say I'm not clever enough to simulate OLG. Could I just choose some crazy high value for the feedback resistor?
I opened an Esnips acount. I hope you can follow the links below.TzeYang said:
http://www.esnips.com/web/Foldedcascodeheadphoneamplifier/. Please tell me what you're working on at the moment?
I hope the RF-filter affects the sound less than the RF itself. I had to drop the switch to make room for it.Sigurd Ruschkow said:
My source has built in caps.
There are really only two pots/channel. One for volume and one for DC offset. The RC/RD are just degenerating (I think this is what they're called) resistors.
I know JFET's are very temperature sensitive, but I didn't know it was crucial to have a stable CCS for the LTP in a folded cascode amp. Is it just as important to have a stable CCS for the cascode? What CCS would you suggest?roender said:
I'm happy with LTSpice or trial and error.TzeYang said:
I will, thank you. I hope I understand it all. It's too bad mr Wurcer and the guys at Analog hasn't written a nice easy-to-understand paper about it, since they use this topology so often with good results.roender said:
So this is the modified circuit with RF filter on the input, Zoebel network on the output and resistors to isolate the rail-to-rail caps(not shown in the schematic, I was thinking something like 0.2 Ohm or so). Everthing is still star grounded. There's still some room to improve the LTP CCS so please make suggestions for it.
You have to click the links this time, I'm afraid. I hope they are visible in Malaysia as well.
http://www.esnips.com/doc/86de8630-59be-4887-99ba-5a6392256f5c/RF-filtered-folded-cascode
http://www.esnips.com/doc/43fc0893-410c-4efb-8f50-6e226f4a314f/RF-filtered-folded-cascode-lay
http://www.esnips.com/doc/9eda9084-1ae6-4bfb-89ab-7b44b32b0e2b/RF-filtered-folded-cascode-top
http://www.esnips.com/doc/68c49a34-5d49-4aa7-8741-198bb6627bd9/RF-filtered-folded-cascode-bottom
Hi,
the RF filter will not affect sonics as long as you use good quality parts for it. Of course you should use good quality parts in the whole amplifier (for best sonics).
As your amp will be a portable amp, it will be exposed to lots of various noise elements like cell phones and motors. I think it is even more important to have a good input filter when an amp is portable and used carried around.
Ah, I see, there is no pot in the LTP. Sorry for the mixup.
There are no large currents so the resistors that separate the channels could be a few Ohms or so.
To simulate OL, you can try your idea.
Another way is to use a huge cap (Gigafarads for ex) in parallel to the RR resistor (330 Ohms).
Sigurd
the RF filter will not affect sonics as long as you use good quality parts for it. Of course you should use good quality parts in the whole amplifier (for best sonics).
As your amp will be a portable amp, it will be exposed to lots of various noise elements like cell phones and motors. I think it is even more important to have a good input filter when an amp is portable and used carried around.
Ah, I see, there is no pot in the LTP. Sorry for the mixup.
There are no large currents so the resistors that separate the channels could be a few Ohms or so.
To simulate OL, you can try your idea.
Another way is to use a huge cap (Gigafarads for ex) in parallel to the RR resistor (330 Ohms).
Sigurd
roender said:
nelsonvandal said:
Roender,
you have a good point. Being a portable amp, it will be exposed to some non usual temperature swings. To me a portable amp is one that is carried around on the bus, on the train, on the beach, on the skiing vacation etc etc.
Thinking of this, what performance loss would you excpect due to the thermal issues?
Sigurd
you have a good point. Being a portable amp, it will be exposed to some non usual temperature swings. To me a portable amp is one that is carried around on the bus, on the train, on the beach, on the skiing vacation etc etc.
Thinking of this, what performance loss would you excpect due to the thermal issues?
Sigurd
roender said:
The amp will mostly rest in my back pocket while sorting the laundry, doing the dishes, picking up after the kids etc, or be used as a desktop amp. So I guess temperature will be an issue, the question is how much and what it will do to the sound.
An amp such as this will be too large to carry around outside. When I've finished this one, I'm ready to take the challange to make something really small. I can see two concepts - either a simpler topology (still CCS's and/or current mirrors are needed because of voltage drop) class AB, 9 V supply for "normal" phones or a unity gain/buffer amp for IEM's or other low Z portable phones where amplification is not needed. Gain only causes noise when using low Z, highly efficient, phones. The latter seems more sane, because why use a tiny amp while carry huge cans.
An amp such as this will be too large to carry around outside. When I've finished this one, I'm ready to take the challange to make something really small. I can see two concepts - either a simpler topology (still CCS's and/or current mirrors are needed because of voltage drop) class AB, 9 V supply for "normal" phones or a unity gain/buffer amp for IEM's or other low Z portable phones where amplification is not needed. Gain only causes noise when using low Z, highly efficient, phones. The latter seems more sane, because why use a tiny amp while carry huge cans.
nelsonvandal said:
In LTSwCAD open File->examples->Educational->LoopGain2.asc, copy squared Vi and Ii (voltage and current source) and place them in your headamp circuit in the same manner, meaning you have to brake the feedback loop and place Vi and Ii between headamp output and load/feedback resistor. Vi and Ii will became Vi1 and Ii1 after paste, so just rename them to Vi and Ii again. Then place in your sch this spice directive
.step param prb list -1 1 ; set prb=0 to turn off probe
.ac dec 30 .1 100Meg
and this comment
-1/(1-1/(2*(I(Vi)@1*V(x)@2-V(x)@1*I(Vi)@2)+V(x)@1+I(Vi)@2))
Then all others spice directive mark as comment, copy -1/(1-1/(2*(I(Vi)@1*V(x)@2-V(x)@1*I(Vi)@2)+V(x)@1+I(Vi)@2)) expression, ran the AC analysis, in "select visible waveforms" card ALT+double click to enter expression, Ctrl+M, paste -1/(1-1/(2*(I(Vi)@1*V(x)@2-V(x)@1*I(Vi)@2)+V(x)@1+I(Vi)@2)) expression, press OK, now you should be able to see your open loop/phase response. Now, place your cursor over green -1/(1-1/(2*(I(Vi)@1*V(x)@2-V(x)@1*I(Vi)@2)+V(x)@1+I(Vi)@2)), right click, in attached cursor choose 1st, press OK, move with mouse (press left and hold) vertical line until horizontal one reach 0dB. Now you can see at LoopGain card your Freq., Magnitude, phase and Group Delay. Forgot to say, you input should be grounded while performing this!
... or you can just post your asc file here and I will do it.
I just got my boards from GoldPhoenix in China. They're not the nicest, but OK. Excellent service. Fast delivery, one week from gerbers to my home in Sweden in a week. Cheap, 11 boards including shipment to Sweden was 129 USD. The boards can be seen at the link below.
http://www.esnips.com/web/Foldedcascodeheadphoneamplifier/
Since I only can get 2SK170GR here in Sweden, will it be OK or should I get 2SK170BL from abroad? Is it possible to parallel two GR's?
http://www.esnips.com/web/Foldedcascodeheadphoneamplifier/
Since I only can get 2SK170GR here in Sweden, will it be OK or should I get 2SK170BL from abroad? Is it possible to parallel two GR's?
If you need BL or GR grade depends on what currents you want to have through your JFETs. If you buy 100 pcs of GR grade from for ex www.elfa.se (where I get mine), you will get a good spectrum of Idss values (within the grade range of course) to choose from.
If you need more than 6.5mA, then you need to get BL or V grade from abroad, which is no problem really. Try Ampslab for ex.
Sigurd
If you need more than 6.5mA, then you need to get BL or V grade from abroad, which is no problem really. Try Ampslab for ex.
Sigurd
nelsonvandal said:
ok this is just imho .... but may be important for you too ...
everytime i listen to headphones at pleasurable listening levels, next day i have tinnitus (sp? meaning: high freq. noise in ears) as well as minor ear pain and some hearing insensitivity (esp at high freq) for the first half of the day or so, then it goes away.
when i listen to stereo through loudspeakers, even when i crank it up all the way on 1kW+/ch stereo for extended periods, i wake up next day with none of the above symptoms ... ears feel fine and sense fine.
now my headphone amp is capable of only ~1.2W/ch yet my stereo is capable of >1kW/ch ... beginning to wonder if listening to loudspeakers is healthier despite the enormous differences (~10000dB) in power output than listening to headphones?
my ears tell me this is true. thoughts?
everytime i listen to headphones at pleasurable listening levels, next day i have tinnitus (sp? meaning: high freq. noise in ears) as well as minor ear pain and some hearing insensitivity (esp at high freq) for the first half of the day or so, then it goes away.
when i listen to stereo through loudspeakers, even when i crank it up all the way on 1kW+/ch stereo for extended periods, i wake up next day with none of the above symptoms ... ears feel fine and sense fine.
now my headphone amp is capable of only ~1.2W/ch yet my stereo is capable of >1kW/ch ... beginning to wonder if listening to loudspeakers is healthier despite the enormous differences (~10000dB) in power output than listening to headphones?
my ears tell me this is true. thoughts?
I have tinnitus as well. It has a kind of natural fluctuation. I agree it gets worse when I'm exposed to loud noises, especially low frequencies like in a car or a bus, but also listening to phones. It all depends on listening levels. That's why I've always been advocating moderate listening levels.gain said:
My first discrete headphone amps were just buffers. Even driving low impedance phones I think gain 1 is enough, except for some recordings, that's why some gain is needed. Many people disagree, and they seem to play loud as hell - and never experience tinnitus.
I haven't been really happy with this amp, too soft sounding, and been fooling around a bit. Removing the active ground channel improved things a bit. Chaning the cascode transistors from BC550/560 to SC2240/SA970 made it a bit brighter, but it was still too soft.
Now I've reduced the bias to the cascodes from 4 mA to 1 mA, and it altered the sound a lot. It's brighter, livelier but still clean. Just better.
How can such a small change make such a difference?
Now I've reduced the bias to the cascodes from 4 mA to 1 mA, and it altered the sound a lot. It's brighter, livelier but still clean. Just better.
How can such a small change make such a difference?
Through D3/D4, determined by J11.AndrewT said:
I've also lowered the current through the cascodes by altering RE/RF but it made no (obvious) difference. Changing the 1N4148 diodes (+ RE/RF) to two red LEDs maybe made it a bit fatter, but I think it was a placebo effect and changed back because of the lowered voltage swing.
Attachments
Yes, the bias current to the cascode made all the difference. From a soft uninvolving sound, to a razor sharp sound - for good and for bad. I can't measure it, but I imagine there are very little of low order distortion, giving it a bit thin but very clear sound.
If you've heard ADA4899, AD8066 or LM4562, you know what I mean.
I guess this is how it's expected to sound, what do you say?
If you've heard ADA4899, AD8066 or LM4562, you know what I mean.
I guess this is how it's expected to sound, what do you say?
gain said:
Keep in mind, your 1.2 watts of headphones are strapped to your ears. In all likelihood you listen at higher levels through headphones. When you listen through loudspeakers you feel it in your body as well as hearing it, you don't when wearing headphones and so you turn up the volume to compensate.
Tinnitus, sensitivity, pain, these are symptoms of excessive noise exposure. You're permanently damaging your hearing, for your own sake please turn down your headphones.
When you sit down to listen through headphones, set the volume to a comfortable level right at the beginning of the listening session, don't increase the volume after that.
P.S. Your amplifier power ratings have no direct correlation with playback volume. A 1000w amplifier playing at anything less than continual clipping is not outputting 1000w. You probably average only a few milliwatts through your headphones, and between a few hundred milliwatts and a couple watts through your speakers.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.