Nah... usually 150-175mA for such a value. Measure DCV drop across the 10 Ohm resistor's pins and you will get the exact answer. I=V/R. Also check that L-Adapter link in my previous post. It was developed to address exactly what you are looking for i.e. to replace SMPS bricks with linear power for devices like your miniDSP digital crossovers etc.
Hi Salas,
Is the Mesmerise board the only board available now, or is Hypnotize still available?
I am also considering making my own board as I have very good equivalent components with different pin spacings. Is it possible to get a copy of the PCB layout please? It would be help me to get close to the original layout, to avoid issues.
Why do you not use a snubber over the transformer secondary's ? These would greatly reduce the switching noise getting to the regulators in the first place.
Kind Regards..
Is the Mesmerise board the only board available now, or is Hypnotize still available?
I am also considering making my own board as I have very good equivalent components with different pin spacings. Is it possible to get a copy of the PCB layout please? It would be help me to get close to the original layout, to avoid issues.
Why do you not use a snubber over the transformer secondary's ? These would greatly reduce the switching noise getting to the regulators in the first place.
Kind Regards..
Back when DCB1 was introduced DIY people did not have Quasimodo test ringer and the like. But the result was good as it was and it still is. Don't know, maybe it successfully blends in a bit of subjectively welcome zing. Each transformer model + bridge diodes combo rings differently. I have nothing against adding anti-ringing network if you are geared up to tailor it on the bench for your specific Tx and rectifiers type.
There is the Mezmerize in the DIYstore for sure. The classic Hypnotize is not available. Not sure if Tea-Bag still has few of the FlexSink Hypnotize rare beast boards available but you can PM him. That was my last edition of the Hypnotize theme. Flexible for onboard normal mode sinking or outboard heavy hot rod sinking. Here's a bird's eye view of layout so to aid your efforts.
There is the Mezmerize in the DIYstore for sure. The classic Hypnotize is not available. Not sure if Tea-Bag still has few of the FlexSink Hypnotize rare beast boards available but you can PM him. That was my last edition of the Hypnotize theme. Flexible for onboard normal mode sinking or outboard heavy hot rod sinking. Here's a bird's eye view of layout so to aid your efforts.
Thanks very much Salas. The layout helps a lot. I have a pair of 2SK389BL, as well as a few 170's. Did wonder if the monolithic dual FET is preferable, as besides the closer match of the FETS, it may help with long term drift.
I built the original B1 and tried this with my more recent set up. I was surprised by its nice open but warm presentation. Hence my interest in the DC version. I guess the potential step improvements come through the lack of coupling caps and a very low noise PSU. Would you say the Hot rodded version is a minor step improvement? Laws of diminishing returns and all..
Kind Regards..
I built the original B1 and tried this with my more recent set up. I was surprised by its nice open but warm presentation. Hence my interest in the DC version. I guess the potential step improvements come through the lack of coupling caps and a very low noise PSU. Would you say the Hot rodded version is a minor step improvement? Laws of diminishing returns and all..
Kind Regards..
Having few K170s won't easily produce two closely matched pairs for minimal offset so prefer a monolithic dual package for each channel. Provided that their Idss (hence Yfs too) is relatively close between them. Like no more than 2mA apart. For the channels performance to track well because the JFETs are biased right at Idss in this crkt.
The DCB1 has been built by hundreds of diyers posting reports through the years. Some of them already had a regular B1 to compare. They said its an audible upgrade. When you need a buffer controller the DCB1 as a design total is a DIY classic by now and a no-brainer.
The hot-rod approach is an indulgence that makes the shunt supply rails more impervious to THD modulations by the audio signal. The PSU's current bias is set silly high vs load demand and the big Mosfets open up in transconductance. Subjectively it sounds grander without sacrificing detail. Mainly worth it in systems with big enough loudspeakers in acoustically treated rooms.
The DCB1 has been built by hundreds of diyers posting reports through the years. Some of them already had a regular B1 to compare. They said its an audible upgrade. When you need a buffer controller the DCB1 as a design total is a DIY classic by now and a no-brainer.
The hot-rod approach is an indulgence that makes the shunt supply rails more impervious to THD modulations by the audio signal. The PSU's current bias is set silly high vs load demand and the big Mosfets open up in transconductance. Subjectively it sounds grander without sacrificing detail. Mainly worth it in systems with big enough loudspeakers in acoustically treated rooms.
I just completed a DCB1/Hypnotize build I started many years ago (2011?). I believe it is the "Black Edition".
Unfortunately, there seems to be a problem. The LED triplets for both positive and negative voltage are not glowing (the series of 5 works).
I measured the voltage around Rset (10 Ohms/5W) them ~ 8.4V for both, which indicates a current much higher than the intended 600mA, so I have to switch off the circuit after approximately 30 seconds - before the resistors get too hot.
I was able to perform a few additional measurements:
Vcc to ground ~40V before regulation.
The voltages at Vout are +-10.4V as expected.
It seems that circuit is still capable of regulating the input voltage down to the intended values.
The relais switches on the buffer section. The voltages at the signal outputs with no signal present are fine: 3mV and 5mV.
All the LEDs (are fine / light up when tested with a DVM, they are all oriented coherently + in accordance with the circles/squares on the PCB.
Any ideas or suggestions what the check next?
Unfortunately, there seems to be a problem. The LED triplets for both positive and negative voltage are not glowing (the series of 5 works).
I measured the voltage around Rset (10 Ohms/5W) them ~ 8.4V for both, which indicates a current much higher than the intended 600mA, so I have to switch off the circuit after approximately 30 seconds - before the resistors get too hot.
I was able to perform a few additional measurements:
Vcc to ground ~40V before regulation.
The voltages at Vout are +-10.4V as expected.
It seems that circuit is still capable of regulating the input voltage down to the intended values.
The relais switches on the buffer section. The voltages at the signal outputs with no signal present are fine: 3mV and 5mV.
All the LEDs (are fine / light up when tested with a DVM, they are all oriented coherently + in accordance with the circles/squares on the PCB.
Any ideas or suggestions what the check next?
The main CCSs don't current limit for some reason. Since you checked its not due to the LED triplets (verify orientation against the theoretical schematic too), see if the associated nearby MOSFETs are the correct N or P type and not accidentally swapped for positive and negative side.
Also, what AC secondary value transformer you use that goes at 40V rectified? Seems high for the application if you mean one side raw DC rail measured and not across rails.
Also, what AC secondary value transformer you use that goes at 40V rectified? Seems high for the application if you mean one side raw DC rail measured and not across rails.
Thanks a lot. I finally found some time to check the orientation again.
The triplets were indeed not oriented correctly. I oriented all the LED according to the hole pattern (square, circle), while it seems that the "D" symbols indicate the correct orientation. The DCB1 one has been working flawlessly for some weeks now and sounds great.
Do you think, changing to 3,3 Ohms would be a significant improvement?
The triplets were indeed not oriented correctly. I oriented all the LED according to the hole pattern (square, circle), while it seems that the "D" symbols indicate the correct orientation. The DCB1 one has been working flawlessly for some weeks now and sounds great.
Do you think, changing to 3,3 Ohms would be a significant improvement?
Congratulations for fixing it. Nice to know that you like it.
Depending on sinks size you could try 3.3 Ω and judge for yourself how you prefer in the concept of your own system.
More PSU bias current does generally differentiate the sound of the DCB1.
Depending on sinks size you could try 3.3 Ω and judge for yourself how you prefer in the concept of your own system.
More PSU bias current does generally differentiate the sound of the DCB1.
Anyone can tell the correct power transformer to feed as it should, the DCB1?
I've got a 65VA R-Core 15-0-15 2.1A, but it runs somewhat hot.
Thanks a lot!
I've got a 65VA R-Core 15-0-15 2.1A, but it runs somewhat hot.
Thanks a lot!
Doesn't seem to be a problem. How many mA shunt reg psu bias you have set? Voltage drop across set resistor divided by its Ohmic value answers that.
Sounds normal to me. Some winding companys use less turns and prefer thicker wire. The transformer then might be able to deliver more current due to less resistance, but it also generally runs hotter. More turns with thinner wire means that it can deliver a bit lesser current but it also runs a bit cooler and are less prone to vibrate. The above is a little generalized and assumes the same core, the same amount of copper wire in weight and the same ratio between primary and secondary. And the same load of course.
Maybe!? More turns excite the magnetic flux in the core better than fewer. The amount of turns needed on the primary depends on the VAC. Lesser voltage needs lesser turns I beleave, and thats where the choice comes for the maker.
I've purchased the board here from a DIY member. It seems that have 2x 47 Ohms to bias de PSU.
