In the market to buy a CCS kit for a SE front end. I have two choices, the K and K Audio Cascode Current Source Kit and the DIY Hifisupply CCS kit.
The KandK is based on cascodes depletion Mosfet, whilst the DIY HFS is based on the 10M45S.
Can anyone comment which performs better in terms of output noise and stability?
The KandK is based on cascodes depletion Mosfet, whilst the DIY HFS is based on the 10M45S.
Can anyone comment which performs better in terms of output noise and stability?
Geek is right on. An excellent CCS is about two transistors and three resistors. Why buy a kit? This is maybe, possibly, $4 worth of parts.
http://www.classicvalve.ca/tubepcb.html
An externally hosted image should be here but it was not working when we last tested it.
The link to Greggs page has some schematics:
http://www.classicvalve.ca/docs/CVD_CCS_preliminary_docs.pdf
http://www.classicvalve.ca/docs/CVD_CCS_preliminary_docs.pdf
Hi SY,
Thanks. May I ask, how does it sound ?
Source resistor would that be R3 ?
Have you compare it to choke loading ?
Many thanks again
Thanks. May I ask, how does it sound ?
Source resistor would that be R3 ?
Have you compare it to choke loading ?
Many thanks again
Constant current is really no fun to listen to, it sounds dull. Music is much nicer to listen to than a constant current.
Compared to choke loading, it's no comparison- the CCS has far better LF performance, lower stray reactances, less nonlinearity, less problem with stray fields and unwanted coupling, no hysteresis... oh, yes, and it costs and weighs a fraction of the choke.
The advantage of a choke is when you need a really large voltage swing and have limited B+.
R1 is the source resistor- the two 300R are gate-stoppers. Their value is less critical than their placement, which should be right at the gates.
Compared to choke loading, it's no comparison- the CCS has far better LF performance, lower stray reactances, less nonlinearity, less problem with stray fields and unwanted coupling, no hysteresis... oh, yes, and it costs and weighs a fraction of the choke.
The advantage of a choke is when you need a really large voltage swing and have limited B+.
R1 is the source resistor- the two 300R are gate-stoppers. Their value is less critical than their placement, which should be right at the gates.
Found this link>
http://www.pacifier.com/~gpimm/ccs_performance.htm
It appears the cascoded Supertex DN2540N5 is almost as good as the high-end semi-discrete CCS.
http://www.pacifier.com/~gpimm/ccs_performance.htm
It appears the cascoded Supertex DN2540N5 is almost as good as the high-end semi-discrete CCS.
SY said:
hey-Hey!!!,
I suggest running bigger gate stoppers...620r of the 1/4 Watt Xicon carbon comp from Mouser. Also suggest using DN3445N3 for the bottom one, they're cheaper and smaller. Also consider battery bias, just like the bottom of Gary's enhancement-mode designs. It allows much larger current setting resistance( that the MOSFET is amplifying ). With them determining the R-set value is much simpler too; assume a -1.6V source-gate voltage plus n*3.3V for how ever many batteries you choose. Also use the RC filter on the batteries and add cathode-cathode 18V Zeners to protect the g-s voltage limits. BZX55C18 is a good one.
cheers,
Douglas
Don´t forget to take Wavebourn/Salas discrete sand-choke in consideration as an alternative to CCS.
http://www.diyaudio.com/forums/showthread.php?postid=1796532#post1796532
http://www.diyaudio.com/forums/showthread.php?postid=1796532#post1796532
Hi eeyore,
Back to your original question; from looking at the websites, the DIY HIFI Supply CCS is the same circuit as the basic K&K Audio CCS. They both could use either the 10M45S or DN2540, which are similar devices and you can expect the same performance from either, although they do sound slightly different. Some folks prefer one, others prefer the other.
K&K Audio also offers a cascode version which will give better performance.
The K&K Audio CCSs also have provisions for a heatsink which is needed if you will dissipate more than 1.5Watts. Both the DN2540 and 10M45S are pretty un-forgiving to overheating so figure your power dissipation and heatsink accordingly.
Dave
Back to your original question; from looking at the websites, the DIY HIFI Supply CCS is the same circuit as the basic K&K Audio CCS. They both could use either the 10M45S or DN2540, which are similar devices and you can expect the same performance from either, although they do sound slightly different. Some folks prefer one, others prefer the other.
K&K Audio also offers a cascode version which will give better performance.
The K&K Audio CCSs also have provisions for a heatsink which is needed if you will dissipate more than 1.5Watts. Both the DN2540 and 10M45S are pretty un-forgiving to overheating so figure your power dissipation and heatsink accordingly.
Dave
David Davenport said:
For a 10 mA current, I'd also consider mixing the 10M45S and a DN3545N3( 10M45S on top ). The g-s voltage of the 10M45S will be ~3V compared to ~1.7 for the Supertex MOSFET. The additional d-s voltage across the lower chip will help reduce gate capacitance. In any case the cascode in any combination will work( and sound ) better than a single.
cheers,
Douglas
David Davenport said:
Anytime! IIRC it was your writing that first brought my attention to the Supertex chips...can't thank you enough for that. Ever so much easier to execute than the enhancement-mode types.
cheers,
Douglas
Hi Baldersnatch,
Thanks for the kind words, and you are welcome. I can claim only to have helped popularize the circuit; the information came from Supertex application notes.
Dave
Thanks for the kind words, and you are welcome. I can claim only to have helped popularize the circuit; the information came from Supertex application notes.
Dave
In the latest AudioXpress, there's a letter from Walt Jung where he shows the performance of a simple DN2540 cascode, basically the same circuit used in the ImPasse and in my upcoming phono preamp article. It is so ridiculously good that I'm unlikely to do any further experimentation down those lines.
Wow, we've come a long way since the Ike Eisenson days of using CR series CC diodes...
Wow, we've come a long way since the Ike Eisenson days of using CR series CC diodes...
- Status
- Not open for further replies.