Hi Nelson,
Thank you again for your comments.
I had thought of the cascoding idea but was worried about changing the sound of the amp by introducing complication and potential interactions. The reason I had originally asked the question about the differences in sound between versions 1 and 2 was because Version 1 looked like a cleaner place to start if I was going to experiment with cascodes.
I am going to experiment by making R6 (and R7) into a divider with the center point driving the base/gate of a cascode. I am intrigued with the idea of using another TL431 as the lower leg of the divider. I know it doesn't sound very minimalist in terms of parts and cost but it looks like it may minimalize the interactions with the rest of the circuit. It's time to build a test bed proto.
Graeme
Thank you again for your comments.
I had thought of the cascoding idea but was worried about changing the sound of the amp by introducing complication and potential interactions. The reason I had originally asked the question about the differences in sound between versions 1 and 2 was because Version 1 looked like a cleaner place to start if I was going to experiment with cascodes.
I am going to experiment by making R6 (and R7) into a divider with the center point driving the base/gate of a cascode. I am intrigued with the idea of using another TL431 as the lower leg of the divider. I know it doesn't sound very minimalist in terms of parts and cost but it looks like it may minimalize the interactions with the rest of the circuit. It's time to build a test bed proto.
Graeme
anyone tried Chokys headphone F4?
Hi, I have built this
http://www.diyaudio.com/forums/showthread.php?postid=1324533#post1324533
Chokys headphone amp and I cant get it to bias at all. I allways triple check components before I solder them and follow the tracks to be sure its correctly label, so I think everything is fine.
I only have 3.6V across my zeners, +/-17V at the drains of the fets. The fets are not matched BTW, will do that later if I think its worth it.
There is also no dc offset. Both pots have been turned full either way and nothing happens. Im using the group buy boards from a while ago, the little ones with the 6 fets mounted either side.
Thought I would post this and go have another look while you all mull it over
Im using j74 k170 jfets and the Vref voltage is -1.5V exactly.
Wonder if my zeners are wrong?
Hi, I have built this
http://www.diyaudio.com/forums/showthread.php?postid=1324533#post1324533
Chokys headphone amp and I cant get it to bias at all. I allways triple check components before I solder them and follow the tracks to be sure its correctly label, so I think everything is fine.
I only have 3.6V across my zeners, +/-17V at the drains of the fets. The fets are not matched BTW, will do that later if I think its worth it.
There is also no dc offset. Both pots have been turned full either way and nothing happens. Im using the group buy boards from a while ago, the little ones with the 6 fets mounted either side.
Thought I would post this and go have another look while you all mull it over
Im using j74 k170 jfets and the Vref voltage is -1.5V exactly.
Wonder if my zeners are wrong?
I changed R9 to 6k8 and it biases now. I do however get very little gain, the sound is very faint. I guess I could listen to them in very quiet rooms
I have a Marantz cd67 with the volume up full. I also have hum. Im going to check my grounding to see what the hum is about, but I would have thought that a cd players 2v out is enough to get my 32ohm cans singing?
Just noticed something peculiar, when I plug in my cd player the bias goes from 300mV to 130mV, surely thats not right?
I have a Marantz cd67 with the volume up full. I also have hum. Im going to check my grounding to see what the hum is about, but I would have thought that a cd players 2v out is enough to get my 32ohm cans singing?
Just noticed something peculiar, when I plug in my cd player the bias goes from 300mV to 130mV, surely thats not right?
Luke said:
I understand that D1 and D2 are there to take the role of voltage-limiters, protecting Vgs and the "six" output MOSFETs safe.
For my F4, I increased R8 value from 22K to 24K instead of reducing R9 value. I had to do this for me to get the bias I wanted. The R8 value of 24K was concluded in consideration of the "actual" reference voltage of TL431 and the "actual" Vgs(threshold) of the output MOSFETs which I used for my F4.
Cheers,
Yeah Luke,
"when I plug in the cd player (assume it's the signal IC wire) and you O/P device bias goes down by 60%" , then Yes , you have a problem! (altho' not with the O/P stage itself!)
This sounds like a wiring problem - maybe in the Gnd/0v connections.
Until you have sorted the problem, suggest add a temporary 100R in series with output - treat the good phones carefully (32R = Grados?)
No volume? It is a buffer, so the cd player mayn't be up to the job, altho' ....
- suggest you check that the 1N4148, 4736 diode are inserted correctly and not cooked (easy mistake, that one)
- assuming the 47k5 input resistor isn't a 4k75
"when I plug in the cd player (assume it's the signal IC wire) and you O/P device bias goes down by 60%" , then Yes , you have a problem! (altho' not with the O/P stage itself!)
This sounds like a wiring problem - maybe in the Gnd/0v connections.
Until you have sorted the problem, suggest add a temporary 100R in series with output - treat the good phones carefully (32R = Grados?)
No volume? It is a buffer, so the cd player mayn't be up to the job, altho' ....
- suggest you check that the 1N4148, 4736 diode are inserted correctly and not cooked (easy mistake, that one)
- assuming the 47k5 input resistor isn't a 4k75
Hi James, thanks for the tips, I havent had time today but I will check my grounding tonight.
Its strange, such a simple little amp The F4 has a buffer so I think the cd player should be able to drive it fine.
My cans are:
http://www.goodgearguide.com.au/index.php/taxid;2136212603;pid;1283;pt;1#cb
These were a bargain and Im quite happy with them, the impeadance on the specs tab says 32ohm, is that unusual for headphones?
Its strange, such a simple little amp
The F4 has a buffer so I think the cd player should be able to drive it fine.My cans are:
http://www.goodgearguide.com.au/index.php/taxid;2136212603;pid;1283;pt;1#cb
These were a bargain and Im quite happy with them, the impeadance on the specs tab says 32ohm, is that unusual for headphones?
AD700 - 32 ohm
The AD700s - quite bright, yes? - rather good with the Cayin hybrid for jazz/classics - the F4 should balance really well with the Silmics for C1,2 - the PanaFCs may be a bit thin.
Needs a good suppy for best results - suggest Mundorfs, Siemens GC, perhaps Roe or Sprague, etc (the heavier sound) & use BYV29 diodes (Rockby Electronics) not cheap diodes or blocks.
The F4 may need extra o/p loading resistors, start with 30R and work up to maybe 150R to find the "sweet spot".
There's quite a few phones with 32R - Grado in particular - can be a problem for lot of h.amps, not this one.
The AD700s - quite bright, yes? - rather good with the Cayin hybrid for jazz/classics - the F4 should balance really well with the Silmics for C1,2 - the PanaFCs may be a bit thin.
Needs a good suppy for best results - suggest Mundorfs, Siemens GC, perhaps Roe or Sprague, etc (the heavier sound) & use BYV29 diodes (Rockby Electronics) not cheap diodes or blocks.
The F4 may need extra o/p loading resistors, start with 30R and work up to maybe 150R to find the "sweet spot".
There's quite a few phones with 32R - Grado in particular - can be a problem for lot of h.amps, not this one.
Hi James,
thanks for the tip, its quite obvious thinking about it, what else could it be, R1 and R2 were swapped round the wrong way. I have one channel going. To be honest I like them but havent heard alot of headphones and havent listened to these much either. Im not sure I find them bright. I have Panasonic FCs throughout, I bought a bunch of them from Dick Smith, but hes dropping components soon and they have no more.
What I have is a diecast box for PSU (big ones from jaycars) and one for the amp with big heatsinks bolted to it, I havent made the case yet, but the plan is to trial all the pass amps out using the same psu, next will be an f5 with the same psu. Its easy to build a class headphone amp. One day when Im bored of trying new stuff I will upgrade diodes and caps etc.
again thanks for your help.
Babowana, thanks for your post, I tried lowering the R9 to 6k8 and it worked fine. But as you can see I had other issues
Channel 2 should be going soon. Only thing better than on healthy channel is two!
thanks for the tip, its quite obvious thinking about it, what else could it be, R1 and R2 were swapped round the wrong way. I have one channel going. To be honest I like them but havent heard alot of headphones and havent listened to these much either. Im not sure I find them bright. I have Panasonic FCs throughout, I bought a bunch of them from Dick Smith, but hes dropping components soon and they have no more.
What I have is a diecast box for PSU (big ones from jaycars) and one for the amp with big heatsinks bolted to it, I havent made the case yet, but the plan is to trial all the pass amps out using the same psu, next will be an f5 with the same psu. Its easy to build a class headphone amp. One day when Im bored of trying new stuff I will upgrade diodes and caps etc.
again thanks for your help.
Babowana, thanks for your post, I tried lowering the R9 to 6k8 and it worked fine. But as you can see I had other issues
Channel 2 should be going soon. Only thing better than on healthy channel is two!
I got both channels going last night, I like it. I still have some weird earthing problem, If I ground both inputs one channel fails, if I earth the one channel both channels go, I f I earth neither inputs lots of hum, im telling myself its a feature for now and will hopefully work itself out when I put it in a box and do proper wiring.
Big thanks to Choky and Nelson, love your work guys.
Big thanks to Choky and Nelson, love your work guys.
Hi Graeme,
The typical resistor value for the currentsource is 220ohm. The current then is 9 (zener voltage) - 4 (Ugsp) = ~23 mA.
To get 23mA with a 6.2 V zener the resistor is (6,2-4)/23 K
=~100ohm.
You will heve a bigger variance in current as the Ugs can vary from 3,7 to 4,4 V (25 - 18mA)
If you use the 6V zeners for gate protection, you will have less reserve however it will still work. Also the maximum output current will be less.
When using the 6V zeners in the power supply, that will work as well. Just use 50% more zeners.
The typical resistor value for the currentsource is 220ohm. The current then is 9 (zener voltage) - 4 (Ugsp) = ~23 mA.
To get 23mA with a 6.2 V zener the resistor is (6,2-4)/23 K
=~100ohm.
You will heve a bigger variance in current as the Ugs can vary from 3,7 to 4,4 V (25 - 18mA)
If you use the 6V zeners for gate protection, you will have less reserve however it will still work. Also the maximum output current will be less.
When using the 6V zeners in the power supply, that will work as well. Just use 50% more zeners.
In my interpretation of the schematic, D1 and D2 are there only for protection and possibly for stability during start up. The constant current (bias) is controlled by Q11 and as long as D1 and D2 are large enough to accommodate Vgs (usually 4.5 V), source resistor voltage drop (0.25V) and a little headroom, you should have no problem using other values.
Thank you both for your responses. I have done similar calculations and come to the same conclusions. It's just that the 6.8V zeners used in the output section of the F4 are significantly lower than the 9.1V parts that Nelson has used in the past. Perhaps there is some other mechanism at work here that isn't obvious. I was asking Nelson the question just to make sure that I wasn't going to run into some obscure interaction.
Graeme
P.S. - Christian, it was great meeting both you and your fiance at BAF . I hope you both enjoyed your visit to San Francisco.
Graeme
P.S. - Christian, it was great meeting both you and your fiance at BAF . I hope you both enjoyed your visit to San Francisco.