Russ, Rudi,
I think what Rudi wanted to say originally with "less noise" was to have a better PSRR. So, given that You are in the fiddling mood, maybe it could be tried as well to fit in, instead of the resistor feed, an another CCS.. Because of the higher dissipation, You could do it with an LM317 as a first approach, or a more elaborated LED-BJT CCS?
[On top of the zeners, to be clear]
It's just an idea, and I am not even sure, because Mauro has good points about the simple Zener shunt.
Maybe you even have tried it already? At least it would clarify if Rudi had some valid point in the better PS separation?
Ciao, George
I think what Rudi wanted to say originally with "less noise" was to have a better PSRR. So, given that You are in the fiddling mood, maybe it could be tried as well to fit in, instead of the resistor feed, an another CCS.. Because of the higher dissipation, You could do it with an LM317 as a first approach, or a more elaborated LED-BJT CCS?
[On top of the zeners, to be clear]
It's just an idea, and I am not even sure, because Mauro has good points about the simple Zener shunt.
Maybe you even have tried it already? At least it would clarify if Rudi had some valid point in the better PS separation?
Ciao, George
Active Current Source
Russ,
I have biased an opamp before using a pair of 2SK170's and a reisistor. How did you impliment it?
Going to the 18 volt zeners sounds like it should be my ticket also. Less heat off the 1K resistors to boot.
George
Russ White said:Rudi,
I think you hit on something.
Increasing the supply voltage of the LM318 also increases its slew rate. I was reading the datasheet to see what sane values would be and decided to perform some experiments.
I actually like the shunt regulators better than the LM78xx/LM79xx as they are actually more quiet(noise wise, if you use a good quiet resistor) and let less AC ripple through to the LM318 (check it with a scope, if you have good zeners you will see there is less ripple).
So what I did, was I took two 18V zeners and popped em in, and then biased the LM318 class A with 7ma with an adjustable current source. I also tried a resistor for kick(3.3K as suggested) but the current source is better.
All I have to say is wow.
This amp has always been excellent, but now I think I am in heaven.
This is on my REV C monoblocs with no other mods at all.
Now keep in mind I have a 300Va toroid on each channel, so my supplies are stiff and clean.
Cheers!
Russ
Russ,
I have biased an opamp before using a pair of 2SK170's and a reisistor. How did you impliment it?
Going to the 18 volt zeners sounds like it should be my ticket also. Less heat off the 1K resistors to boot.
George
Some other mods
Good to hear that the Rev C gets better by the day (or is that by the hour?)
Some other mods that are worth considering include:
1. Discrete diodes, hi-speed, soft recovery of course. These normally result in a lower noise floor when replacing diode bridges
2. 100R in series with 330nF caps as snubbers across the diodes. Same effect as 1.
3. Double regulation for the 318's power supplies? This requires more work, but could be worthwhile based on experiences with NAIM pre-amps
4. Speaking of NAIM, I have always had excellent results from using oversized toroids to provide power supply. I know that Rudi is using relatively small transformers (20VA I think) and it would be easy enough to determine if using say a 100VA makes a worthwhile difference. I would expect it to result in further big improvements in bass resolution.
5. Now that the power supply voltage is at a decent level, it may be worth trying the 318 replacements (627, etc) again, as perhaps these made no difference in the past vs 318 because they were not receiving sufficient juice....
Thats about it from me for now. Anyone else with some ideas on pushing the boundaries?
Ryan
Good to hear that the Rev C gets better by the day (or is that by the hour?)
Some other mods that are worth considering include:
1. Discrete diodes, hi-speed, soft recovery of course. These normally result in a lower noise floor when replacing diode bridges
2. 100R in series with 330nF caps as snubbers across the diodes. Same effect as 1.
3. Double regulation for the 318's power supplies? This requires more work, but could be worthwhile based on experiences with NAIM pre-amps
4. Speaking of NAIM, I have always had excellent results from using oversized toroids to provide power supply. I know that Rudi is using relatively small transformers (20VA I think) and it would be easy enough to determine if using say a 100VA makes a worthwhile difference. I would expect it to result in further big improvements in bass resolution.
5. Now that the power supply voltage is at a decent level, it may be worth trying the 318 replacements (627, etc) again, as perhaps these made no difference in the past vs 318 because they were not receiving sufficient juice....
Thats about it from me for now. Anyone else with some ideas on pushing the boundaries?
Ryan
A note of caution folks. The mods I made have worked great for me, but I did not design this amp, and I am not sure what long term impact the changes I made have on stability or durability and such, so I would not necessarily recommend everyone go and do what I did. If you do it be very careful. I would also like to say that nobody should feel like they will have a substandard amp if they don't modify it. Nothing could be farther from the truth. Seriously just ask the many who have built from Mauro's, Plovati's, or my PCBs and you will see that the My-REF amp as designed is simply one of the best available chipamps period.
Cheers!
Russ
Cheers!
Russ
well i promised a friend that will build him a all out amp.
2 x 500VA C-core (good old fashion over kill)
2 x dual 18v 40 VA transformers.
finder relay's
caps to be decided
silver hookup wire.
he is supplying all the bits
seeing that the current configuration is the second best i have ever heard i think it justifies the expence.
just for the fun of it . he is 75years old needs something to make him stop complaining about the amps out there.
this should be finished by the end of next week
2 x 500VA C-core (good old fashion over kill)
2 x dual 18v 40 VA transformers.
finder relay's
caps to be decided
silver hookup wire.
he is supplying all the bits
seeing that the current configuration is the second best i have ever heard i think it justifies the expence.
just for the fun of it . he is 75years old needs something to make him stop complaining
about the amps out there. this should be finished by the end of next week
Byrd said:i was going to say Russ. 300VA
I will be bringing mine up to test and burn in with 2 100VA toroids (1 per channel) but plan on replacing that with 2 250VA units later on.
The test units are 25 volt (+1 over spec'ed shouldn't hurt the amp) but will likely go to 24 on the final set.
So the qustion is...
What is the optimal size recommended @ 24 volts? 200, 250, 330? I prefer not to WAY over do it.
Ja nee (yes no)
Typical Gauteng move there...
Use a 3 litre Ford to go get a 2L Coke and a 1L Brandy.
You go boet. Proudly South African and all that.
Here it is the night before christmas. Had confirmation today that my specialy wound transformer will be delivered in the morning, lol, also a bit of overkill rated at 6 Amps... and a small 5V 1A winding for leds and small stuff, only cost about R10 extra.
Tomorrow, tomorrow, I love you tomorrow....
Typical Gauteng move there...
Use a 3 litre Ford to go get a 2L Coke and a 1L Brandy.
You go boet. Proudly South African and all that.
Here it is the night before christmas. Had confirmation today that my specialy wound transformer will be delivered in the morning, lol, also a bit of overkill rated at 6 Amps... and a small 5V 1A winding for leds and small stuff, only cost about R10 extra.
Tomorrow, tomorrow, I love you tomorrow....