AVR300, I am having a hard time grasping what it is exactly you are trying to do.
The current through the top half of the legato has to be adjusted for two factors. The positive rail and the I/V resistor.
For 1-1.2V 100R or 105R would work fine for the I/V resistor. You don't need parallel resistors here at all. The resistor will only be dissipating about 1/4W.
Lets assume 15V rails.
Then you will want something like 340-390R at the CCS end (the bottom half of the cct). You can get this value with a single resistor (about 3/4W dissipated so 1W part is fine) or with a couple resistor in parallel. Say 680R.
Increasing the number of FETs does not really come into play except that the transconductance gets better as you add them with the point of diminishing returns coming at about 3.
I hope that helps.
The current through the top half of the legato has to be adjusted for two factors. The positive rail and the I/V resistor.
For 1-1.2V 100R or 105R would work fine for the I/V resistor. You don't need parallel resistors here at all. The resistor will only be dissipating about 1/4W.
Lets assume 15V rails.
Then you will want something like 340-390R at the CCS end (the bottom half of the cct). You can get this value with a single resistor (about 3/4W dissipated so 1W part is fine) or with a couple resistor in parallel. Say 680R.
Increasing the number of FETs does not really come into play except that the transconductance gets better as you add them with the point of diminishing returns coming at about 3.
I hope that helps.
Last edited:
Seeking the lost sound 
and learning how this circuit works. Sorry for being a PITA.
Ok, now, I'm following your advice and are aiming for Vcc/2-1 at output.
With 180R and 390R and +-15v I get 8v at output. I have to trim the Placid to +14 -15 to obtain 6.5v at output. This is strict opposite the advice to trim the negative side of the psu lower than the positive side - and it means ??? (the CSS resistor is too big in value ??)
Ok regarding the FET's, it's analog to my own findings.
Only learning here, hope it's ok.
and learning how this circuit works. Sorry for being a PITA.Ok, now, I'm following your advice and are aiming for Vcc/2-1 at output.
With 180R and 390R and +-15v I get 8v at output. I have to trim the Placid to +14 -15 to obtain 6.5v at output. This is strict opposite the advice to trim the negative side of the psu lower than the positive side - and it means ??? (the CSS resistor is too big in value ??)
Ok regarding the FET's, it's analog to my own findings.
Only learning here, hope it's ok.
In a 400mA environment, could your BS250s have been exhibiting thermal effects (or damaged)? It does seem that there is a short 'warm up' period for my Legato ver. 1... (fairly minor effect but noticeable)
http://www.vishay.com/docs/70209/70209.pdf
AVR,
I am very happy to help people understand the principles involved here.
(VCC/2)-1 bias at the I/V resistor is a rough guideline. No need to try to hit it exactly. With 15V rail anything from 8 to 6 volts there should produce excellent results.
Your 390R at the CCS end could be lowered a bit to get you closer. Say 320-350R.
Your trimming the positive rail is not really a problem, but as a rule the higher it is the better performance will be. So it better to adjust either the negative rail or the CCS R to get your I/V bias voltage to where you want it.
I am very happy to help people understand the principles involved here.
(VCC/2)-1 bias at the I/V resistor is a rough guideline. No need to try to hit it exactly. With 15V rail anything from 8 to 6 volts there should produce excellent results.
Your 390R at the CCS end could be lowered a bit to get you closer. Say 320-350R.
Your trimming the positive rail is not really a problem, but as a rule the higher it is the better performance will be. So it better to adjust either the negative rail or the CCS R to get your I/V bias voltage to where you want it.
Last edited:
Russ, it was suggested over on pfm that the coupling caps (nichicons) are not required if solely using legato for SE output. Can you confirm if this is ok to run this way please? I currently have mine setup like this and it sounds much much better than with any cap in place. Just want to make sure nothing's going to fry.
It is possible to run without caps that way as long as the SE offset is low enough. This will not always be the case. If your measured SE offset is low then you are fine. You will just need to check that carefully. If you like you can even trim out the offset using a parallel trim resistor on the CCS for the side that you want to adjust downward.
If you only use SE output you can definitely use lower value film caps. Say 2.2-4.7uf. Then you will be completely safe and no worries about offset in the least.
If you only use SE output you can definitely use lower value film caps. Say 2.2-4.7uf. Then you will be completely safe and no worries about offset in the least.
Last edited:
You put me on the track again. The 400mA was burned in the power supply. I newer measured the actually load.
Along with changing the resistors, I trimmed down the Placid to around 250mA. That would leave 50mA in the shunt. And I did not have voltage drops.
Today, I tried this. Burning 300mA, leaving the 100mA in the shunt itself. It gets warmer, but it's not critical.
Guess what, it sounds much better now.
That's great news AVR! When I first checked R13 and R14 on the Placid BP they were just over 100 mV (and Legato was sounding wonderful). Out of caution I trimmed back to 80 mA overhead and still extra good sound. The Placid heat sinks seem to manage easily (free standing), though Russ's advice about adequate chassis ventilation will surely be important. Meanwhile, I've acquired some new source material - a few 96/24 master tracks - and will confirm the best supply parameters for that music (vs. 44/16 material upsampled in software to 96/24). Skol!
You put me on the track again. The 400mA was burned in the power supply. I newer measured the actually load.
Along with changing the resistors, I trimmed down the Placid to around 250mA. That would leave 50mA in the shunt. And I did not have voltage drops.
Today, I tried this. Burning 300mA, leaving the 100mA in the shunt itself. It gets warmer, but it's not critical.
Guess what, it sounds much better now.
Hi AVR300
Interesting, perhaps the current draw for the Legato is not as constant as believed? - As I mentioned on PFM, I used OS-CONS as the supply decoupling caps on the Legato, I wonder if the board is sensitive to type used? I am shunting 60mA, and the regs are hot but probably could go higher. Interestingly my 5v Placid runs hotter than the 15v.
Of course, when designing my case I was concerned about venting over the Legato, so I may need to provide a few more case holes near the placids now.
Regarsd
Curiously, in several days of use the current overhead of the Placid has crept from (spot on) 80mA up to 98. For now I'll leave it. Could Legato current usage have declined?
As for my newer reference music, I find that a slightly lower Vplus, 14.25v, is pleasing with the extended dynamic range of 96kHz/24bit master tracks. However, that certain 'magic quality' I mentioned previously is diminished when Vminus is also lowered to 14.25, so I'm leaving Vminus at 14.75. I had not previously tried a higher Vminus than Vplus... It's amazing how you can hear stuff like the collective embouchure of a french horn section!
I KNEW you would bust me on that, Avr! Since switching to OS X I haven't used special accents but now you've forced me to learn how. For that...
Sköl!
As for my newer reference music, I find that a slightly lower Vplus, 14.25v, is pleasing with the extended dynamic range of 96kHz/24bit master tracks. However, that certain 'magic quality' I mentioned previously is diminished when Vminus is also lowered to 14.25, so I'm leaving Vminus at 14.75. I had not previously tried a higher Vminus than Vplus... It's amazing how you can hear stuff like the collective embouchure of a french horn section!
ascii 143
I KNEW you would bust me on that, Avr! Since switching to OS X I haven't used special accents but now you've forced me to learn how.
For that...Sköl!
I would suggest this would most likely to be due to the temperature variation on the CCS transistor. The hotter this gets, the more current it will push out. Compare the current variation on both the CCS resistor and the shunt resistor to see if it is the power supply increasing current itself, or the load. If the voltage on both resistors increases the same, it's just the PSU sourcing & sinking more current itself, nothing to do with the legato.
After a week of perfect performance I can now hear very faint 'scratchy' noise from one channel. Power resistors are PRP 1W and running hot w/ 190 mA/side. I'm considering replacing each with pairs mounted below the board rather than above. Certainly the inboard resistors will go below because the Buffalo board above them is taking on more heat than I would like. Photo of my board: https://files.me.com/francolargo/qbapa2
Thanks for the insights AussieAB! ...will monitor both voltages w.r.t. temp.
Frank
Thanks for the insights AussieAB! ...will monitor both voltages w.r.t. temp.
Frank
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.