Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Dirt Cheap Stax Amp DIY - new schematic updated!
New Posts  All Forums:Forum Nav:

Dirt Cheap Stax Amp DIY - new schematic updated! - Page 3

post #31 of 96

All of the stax tube amps do have global feedback. So the high frequency roll off

is compensated for.

 

Better to use constant current sources for the plate loads. Then you end up

with ixys parts and then you decide you need bigger tubes, and then you

decide to reference the outputs to ground, and sooner or later it looks like

a T2.

post #32 of 96

Yes, even the very old Stax amps had feedback.  Only the super simple "A" design they released for DIY use in 1968 was without feedback.  I'm listening to that design now modified with a differential front end and it's ok for something so simple.  Works off a single filament supply too which makes sourcing a transformer that much easier. 

 

The plate resistors are compromise but what else can we do on a budget?  We could use tubes as current sources but then we enter filament hell.

 

One thing I'm going to try soon is run small encapsulated transformers backwards off 24VAC to generate high voltages on the cheap.  18V/230V would give us plenty of voltage to play with and regulating it would be easy. 

post #33 of 96

are we going to keep hearing this FUD about properly selected, properly applied MOSFET right up to the moment you guys finally "discover" how to use them right?

 

just read the datasheet - you do have the one with the full plots?  http://ixapps.ixys.com/DataSheet/DS98809C(IXTP-U-Y01N100D).pdf

 

if you can spare 25 V then 12 pF Coss, 2 pF Crss aren't crippling with ~100 pF headphone load

 

if you draw up the full bridge, headphone load and device parasitics then only ~ 15 % of the current is wasted in the MOSFET parasitic C, 1/2 that if the IXTP01N100 MOSFET are only used for CCS


Edited by jcx - 3/23/13 at 2:54pm
post #34 of 96

Already using those parts in the kgsshv.

 

problem is that they have a dramatic drift with temperature

so you need a good servo to deal with that.

 

10m90s much better, but will not go as low in current.

 

Its not the output capacitance that is the issue with some

of the mosfets, but rather the input capacitance which

can be as high as 700pf. Makes them hard to drive at

20khz.


Edited by kevin gilmore - 3/24/13 at 9:17am
post #35 of 96

servos are cheap - just add to parts count

 

but you can also cascode another current regulating circuit with the depletion mode devices - looks like there is >2 V @10mA to work with with the IXTP01N100

 

plenty for LM334, even with the diode TC comp hack

 

or 2 BJT feedback CCS could reduce I bias drift from Vgs TC ~50x, but still would have BJT Vbe TC

 

could use some 3-terminal regulator chips, even precision refs

 

 

 

the specific part under discussion IXTP01N100, should handle enough current for ES headphones - has small die: Cgs ~ 100 pF not that much more than BJT, only has to be modulated ~ 0.5 V

 

acutally less current required than EL34 triode connected with mu <20 where you need >50 Vgk across 15 pF
 


Edited by jcx - 3/24/13 at 10:07am
post #36 of 96
Thread Starter 
Quote:
Originally Posted by spritzer View Post

Yes, even the very old Stax amps had feedback.  Only the super simple "A" design they released for DIY use in 1968 was without feedback.  I'm listening to that design now modified with a differential front end and it's ok for something so simple.  Works off a single filament supply too which makes sourcing a transformer that much easier. 

The plate resistors are compromise but what else can we do on a budget?  We could use tubes as current sources but then we enter filament hell.

One thing I'm going to try soon is run small encapsulated transformers backwards off 24VAC to generate high voltages on the cheap.  18V/230V would give us plenty of voltage to play with and regulating it would be easy. 

Running little transformers backwards may lead to low efficiency. Since the primary winding has very high DC resistance.
post #37 of 96
Thread Starter 
Quote:
Originally Posted by jcx View Post

are we going to keep hearing this FUD about properly selected, properly applied MOSFET right up to the moment you guys finally "discover" how to use them right?

just read the datasheet - you do have the one with the full plots?  http://ixapps.ixys.com/DataSheet/DS98809C(IXTP-U-Y01N100D).pdf

if you can spare 25 V then 12 pF Coss, 2 pF Crss aren't crippling with ~100 pF headphone load

if you draw up the full bridge, headphone load and device parasitics then only ~ 15 % of the current is wasted in the MOSFET parasitic C, 1/2 that if the IXTP01N100 MOSFET are only used for CCS
This MOSFET looks like a HV jfet. Which can be used as SRPP. Are they expensive?
post #38 of 96
Thread Starter 
I am trying to work on a transistor version SRPP, which is cheap( only using 8 BJT in total) and output impedance is really low for ES phones.

Initial tests shows its hard to bias the upper transistor and its prone to temp/voltage drifts. If any of you worked on tube SRPP in the past, how do you adjust the output to be 1/2 of the power supply? Does tubes not drift at all? I only worked in SE tube amps and my knowledge about tubes are limited.
post #39 of 96

And this is the major problem.

 

One solution is 2sa1968. which is impossible to get these days.

 

Another solution is ixys10m90s which is designed to have sharp

cutoff and thus stable current with temperature. But needs to

run at 5ma minimum.

 

latest kgsshv has a servo because the 01n100d as a current

source in the 3rd stage has temperature drift. Not a cheap

design. Lots of parts.

 

With a ground referenced design and a servo, you can definitely

control the output voltage to 1/2 of B+

post #40 of 96

The thought of SRPP and 'stats just brings nasty flashbacks for me.  Kevin and I came up with the B-10 design which is a fixed version of the RSA A-10 but one could argue that it isn't really SRPP since the load isn't fixed.  I ran the tubes very hard and it's pretty stable but sounds utterly terrible. 

 

Quote:
Originally Posted by wdiabc View Post


Running little transformers backwards may lead to low efficiency. Since the primary winding has very high DC resistance.

 

Yeah, that is the main drawback but I'm going to the local store which has a huge selection of 230V units and pick the best ones.  Hell, I've used some of the same transformers for audio use (tiny Stax adapters) so this has to perform better...  tongue.gif

post #41 of 96

the IXTP01N100D are currently ~$4 ea, need another few bucks of heatsink, hardware

 

not "dirt cheap" if you need 8 for 2 channels, dual stator drive - but could still be less than cost of tubes, sockets, isolated heater supply windings

 

 

I have protoed one stator driver with the IXYS, and "SRPP derived" modulated current source - but used op amp gain inside the global feedback loop so the DC servo was built in so to speak

 

 

too lazy to handwire point-to-point any more than the "proof of concept" - need to brush off the Eagle and layout some test boards

 

especially since I plan to totem pole/cascade 2x each Q for a ESP-950 +/-600 Vsupply amp which doubles the parts count, and all the cool op amps today are smt anyway


Edited by jcx - 3/26/13 at 10:18pm
post #42 of 96
Thread Starter 
Quote:
Originally Posted by jcx View Post

the IXTP01N100D are currently ~$4 ea, need another few bucks of heatsink, hardware

 

not "dirt cheap" if you need 8 for 2 channels, dual stator drive - but could still be less than cost of tubes, sockets, isolated heater supply windings

 

 

I have protoed one stator driver with the IXYS, and "SRPP derived" modulated current source - but used op amp gain inside the global feedback loop so the DC servo was built in so to speak

 

 

too lazy to handwire point-to-point any more than the "proof of concept" - need to brush off the Eagle and layout some test boards

 

especially since I plan to totem pole/cascade 2x each Q for a ESP-950 +/-600 Vsupply amp which doubles the parts count, and all the cool op amps today are smt anyway

Can you share any schematics based on this idea? It sounds very interesting. 

post #43 of 96
Thread Starter 

what do you think of this ? That looks like a MOSFET SRPP with a global feedback, still can be done cheap with OP amps and MOSFETs, I feel IRF730/740 can be acquired pretty cheap. Just don't know how stable this circuit is.

post #44 of 96

Gilmore does have a point if you choose MOSFET that big with too large parasitic C - the IXYS part seems to be very specially suited to ES headphone amps - can be bought at DigiKey

 

you really don't want multi-Amp rated parts for the ~ 10mA you may actually use

 

the BSP125 in that schematic are more reasonably sized for the job - if a bit low V, low case pwr rating and smt so harder to proto with


Edited by jcx - 3/27/13 at 10:19am
post #45 of 96

the schematic posted above is that of the sennheisser hev60.

Most people who have listened to that amp don't like it.

 

Its only SRPP if the load impedance matches the cathode resistor.

In this case the source of the top fet. Therefore its not SRPP but

rather constant current.

 

In fact, due to C20 its more like a white cathode follower done

with solid state.


Edited by kevin gilmore - 3/28/13 at 3:21am
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Dirt Cheap Stax Amp DIY - new schematic updated!