@ikoflexer
it is well known 24/192 DAC from Ebay. I am using it to experiment. This "el-cheapo" is not bad, considering price & after some modifications, of course.
Bad thing is CS4397 get easily burned, but have several chips in my spare parts box. Replacement is 3-minute job.
Shematic is here: http://www.shrani.si/?1X/x5/9esiKsT/dac1.jpg
http://www.shrani.si/?3F/9O/2G6LodNt/boardshema.jpg
@AndrewT
I am better with soldering iron as with computer drawings
I am old fart with shaking arms, my eyes having problems with SMD parts, but I can still read my multimeter with 2,5 cm big digits
Shunt reg at 5V is only for receiver & DAC chip. Load is going from 56.12 mA and 56.16 mA. For me this is constant load.
Output from DAC is going through 3rd order filter to analog/buffer JFET stage, very similar to Borbely's EB604/410 All-Jfet Line Amp and powered from 24V.
Not enough time to finish my Aikido line stage
@anatech
As far as I understand Teddy's Super reg, first stage is LP filter, with pass transistor. First transistor have similar function as first MOSFET in Salas shunt reg.
it is well known 24/192 DAC from Ebay. I am using it to experiment. This "el-cheapo" is not bad, considering price & after some modifications, of course.
Bad thing is CS4397 get easily burned, but have several chips in my spare parts box. Replacement is 3-minute job.
Shematic is here: http://www.shrani.si/?1X/x5/9esiKsT/dac1.jpg
http://www.shrani.si/?3F/9O/2G6LodNt/boardshema.jpg
@AndrewT
I am better with soldering iron as with computer drawings
I am old fart with shaking arms, my eyes having problems with SMD parts, but I can still read my multimeter with 2,5 cm big digits
Shunt reg at 5V is only for receiver & DAC chip. Load is going from 56.12 mA and 56.16 mA. For me this is constant load.
Output from DAC is going through 3rd order filter to analog/buffer JFET stage, very similar to Borbely's EB604/410 All-Jfet Line Amp and powered from 24V.
Not enough time to finish my Aikido line stage
@anatech
As far as I understand Teddy's Super reg, first stage is LP filter, with pass transistor. First transistor have similar function as first MOSFET in Salas shunt reg.
Hi Salas,
I don't know about being faster is an answer or not. Normally a bypass capacitor satisfies those needs well enough. Of course, a bypass capacitor may cause trouble depending on the characteristics of a shunt. Noise in the shunt may translate into high current pulses as the capacitor(s) attempt to maintain the voltage across their contacts.
If you have digital and analog sections on the same supply, it needs to be a stiff as possible, hence the larger values of bypass capacitors, to a reasonable limit of course! The demands on a purely analog section should be a great deal less, but you still have to deal with current demands across the entire audio range, right on up to the sampling frequency. That can be pretty high. Again, you are looking for a voltage source with infinite current available instantly. That leaves you with a choice between using capacitors or a pass element capable of video frequencies. The filter capacitor route may generate less noise in a pass configuration. It's cheaper too!
Personally, good supplies can be constructed either way. You have to imagine the supply from the load's point of view. It doesn't care how the power gets there, it only cares that it is stable and can supply small current pulses as needed. I would guess that a current output DAC will draw less. If you use a voltage output DAC and install small capacitors to ground at the output, you're going to draw some peaks of current. So I would say that the filter design may have more influence than the power supply on the sound quality. That's assuming that the power supply is of good design either way (series or shunt).
-Chris
I don't know about being faster is an answer or not. Normally a bypass capacitor satisfies those needs well enough. Of course, a bypass capacitor may cause trouble depending on the characteristics of a shunt. Noise in the shunt may translate into high current pulses as the capacitor(s) attempt to maintain the voltage across their contacts.
If you have digital and analog sections on the same supply, it needs to be a stiff as possible, hence the larger values of bypass capacitors, to a reasonable limit of course! The demands on a purely analog section should be a great deal less, but you still have to deal with current demands across the entire audio range, right on up to the sampling frequency. That can be pretty high. Again, you are looking for a voltage source with infinite current available instantly. That leaves you with a choice between using capacitors or a pass element capable of video frequencies. The filter capacitor route may generate less noise in a pass configuration. It's cheaper too!
Personally, good supplies can be constructed either way. You have to imagine the supply from the load's point of view. It doesn't care how the power gets there, it only cares that it is stable and can supply small current pulses as needed. I would guess that a current output DAC will draw less. If you use a voltage output DAC and install small capacitors to ground at the output, you're going to draw some peaks of current. So I would say that the filter design may have more influence than the power supply on the sound quality. That's assuming that the power supply is of good design either way (series or shunt).
-Chris
Hi stormsonic,
I do think he touched on a valid problem though. That is the inability, or reduced ability, for normal three terminal regulators to block the higher frequency noise. Mind you, if you don't have to loss some voltage, an L-C filter combination before the regulator might fix that. If you do have to drop some voltage, I'd be using a switching regulator first. Then L-C filters can be extremely effective at removing the high frequency components. Switching regulators switch from about 200 KHz to 1 MHz and beyond these days, and they are not exotic parts.
More than one way to skin a cat, and the Teddy Super reg. will work. If I'm going to go to all the trouble of building something, I want better performance.
-Chris
True, but my point is that there is no correction for IR losses at all. I wouldn't call it an effective voltage regulator.
I do think he touched on a valid problem though. That is the inability, or reduced ability, for normal three terminal regulators to block the higher frequency noise. Mind you, if you don't have to loss some voltage, an L-C filter combination before the regulator might fix that. If you do have to drop some voltage, I'd be using a switching regulator first. Then L-C filters can be extremely effective at removing the high frequency components. Switching regulators switch from about 200 KHz to 1 MHz and beyond these days, and they are not exotic parts.
More than one way to skin a cat, and the Teddy Super reg. will work. If I'm going to go to all the trouble of building something, I want better performance.
-Chris
Hi Tham,
Quick question for you.
When an A/D circuit is changed from an oversampling type, to a non-oversampling type, do you also correct the filter section back to a 7th order type? Either that, or is the DSP or digital filter changed?
I only ask because I figured that since you are experimenting with these, you might know the answer.
One thing I do know is that if you drop the sampling rate, you must also change the anti-aliasing filter in the following analog section. Otherwise you will have quite a bit of ultrasonic noise passing through. Don't expect to hear it.
-Chris
Quick question for you.
When an A/D circuit is changed from an oversampling type, to a non-oversampling type, do you also correct the filter section back to a 7th order type? Either that, or is the DSP or digital filter changed?
I only ask because I figured that since you are experimenting with these, you might know the answer.
One thing I do know is that if you drop the sampling rate, you must also change the anti-aliasing filter in the following analog section. Otherwise you will have quite a bit of ultrasonic noise passing through. Don't expect to hear it.
-Chris
Tham, I've tried quite a few regulators myself, including some Jung variations. Maybe naive, but I was more impressed with salas v1. I cannot recommend v2 because it would be a biased recommendation; people should make their own mind about that.
Chris, hello neighbour! Nice to see you active in this thread.
Edit: typo
Chris, hello neighbour! Nice to see you active in this thread.
Edit: typo
Tham said:
ikoflexer said:tham, thank you for taking the time to build and test v2. You're a brave man
ikoflexer said:
Non the less it has been proven that it works as intended Iko, and we should establish a pcb design for V2 and recommend it as the advanced bandwidth version for people to try as a possible upgrade.
Tham, my higher current v2 +/- is almost finished. 
I got sidetracked with a dht headphone amp design and building a high voltage v2n (v2 with n-channel mosfets) for it. That is almost finished too... if I could only work on one project at a time...
salas, if anyone wants to do the pcb layout, I have nothing against it. As it is, I will not have time to do it right now, plus, my layouts you've seen... not exactly works of art.
so, go for it
I got sidetracked with a dht headphone amp design and building a high voltage v2n (v2 with n-channel mosfets) for it. That is almost finished too... if I could only work on one project at a time... salas, if anyone wants to do the pcb layout, I have nothing against it. As it is, I will not have time to do it right now, plus, my layouts you've seen... not exactly works of art.
so, go for it
stormsonic said:
Post #578. You'll need to change the Vref zeners for your voltage, and the series resistor (R9) for the quiescent current. If you tell me what's your Vout and the idle current you want I can post a schematic with adjusted component values for you to try. The two bjt may need a small cap each to stabilize it (18-200pF range).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.